Method: Uses sparse autoencoders (SAEs) to extract concept activations, computes saliency-weighted performance scores across benchmark data, enabling comparison across benchmarks via model’s internal representations.

Result: Models consistently underperformed on concepts contrasting sycophantic behaviors (politely refusing requests, asserting boundaries) and safety-related concepts. Benchmarks over-represented obedience/authority concepts while missing core concepts within their intended scope.

Conclusion: The method provides concept-level decomposition of benchmark scores, complementing aggregated metrics by revealing why models scored as they did and how benchmarks could better reflect intended scope.

Abstract: The evaluation of large language models (LLMs) relies heavily on standardized benchmarks. These benchmarks provide useful aggregated metrics for a given capability, but those aggregated metrics can obscure (i) particular sub-areas where the LLMs are weak (“model gaps”) and (ii) imbalanced coverage in the benchmarks themselves (“benchmark gaps”). We propose a new method that uses sparse autoencoders (SAEs) to automatically uncover both types of gaps. By extracting SAE concept activations and computing saliency-weighted performance scores across benchmark data, the method grounds evaluation in the model’s internal representations and enables comparison across benchmarks. As examples demonstrating our approach, we applied the method to two popular open-source models and ten benchmarks. We found that these models consistently underperformed on concepts that stand in contrast to sycophantic behaviors (e.g., politely refusing a request or asserting boundaries) and concepts connected to safety discussions. These model gaps align with observations previously surfaced in the literature; our automated, unsupervised method was able to recover them without manual supervision. We also observed benchmark gaps: many of the evaluated benchmarks over-represented concepts related to obedience, authority, or instruction-following, while missing core concepts that should fall within their intended scope. In sum, our method offers a representation-grounded approach to evaluation, enabling concept-level decomposition of benchmark scores. Rather than replacing conventional aggregated metrics, CG complements them by providing a concept-level decomposition that can reveal why a model scored as it did and how benchmarks could evolve to better reflect their intended scope. Code is available at https://competency-gaps.github.io.

Method: SA-DiffuSeq integrates sparse attention into the diffusion framework, selectively allocating attention to reduce computational complexity. It includes a soft absorbing state tailored to sparse attention dynamics to stabilize diffusion trajectories and accelerate sequence reconstruction.

Result: SA-DiffuSeq consistently surpasses state-of-the-art diffusion baselines in both training efficiency and sampling speed, with especially strong gains on extended sequences. It maintains semantic coherence and generation quality while reducing computational overhead.

Conclusion: Incorporating structured sparsity into diffusion models is a promising direction for efficient and expressive long text generation, making SA-DiffuSeq well-suited for demanding applications like scientific writing, code generation, and long-context dialogue.

Abstract: Diffusion based approaches to long form text generation suffer from prohibitive computational cost and memory overhead as sequence length increases. We introduce SA-DiffuSeq, a diffusion framework that integrates sparse attention to fundamentally improve scalability for long document modeling. By selectively allocating attention within the diffusion process, SA-DiffuSeq significantly reduces computational complexity while maintaining semantic coherence and generation quality. A key component of our method is a soft absorbing state tailored to sparse attention dynamics, which stabilizes diffusion trajectories and accelerates sequence reconstruction. This design improves sampling efficiency and enhances precision in long range dependency modeling. Extensive experiments demonstrate that SA-DiffuSeq consistently surpasses state of the art diffusion baselines in both training efficiency and sampling speed, with especially strong gains on extended sequences. These properties make SA-DiffuSeq well suited for demanding long form applications such as scientific writing, large scale code generation, and multi turn long context dialogue. Overall, our results indicate that incorporating structured sparsity into diffusion models is a promising direction for efficient and expressive long text generation.

Method: Created TokSuite with 14 models using different tokenizers but identical architecture, dataset, training budget, and initialization. Also developed a benchmark measuring performance under real-world perturbations affecting tokenization.

Result: TokSuite enables robust decoupling of tokenizer influence, supporting novel findings that elucidate the respective benefits and shortcomings of a wide range of popular tokenizers.

Conclusion: TokSuite provides a systematic framework for studying tokenization’s role in language models, offering insights into how different tokenizers affect model performance and behavior in real-world scenarios.

Abstract: Tokenizers provide the fundamental basis through which text is represented and processed by language models (LMs). Despite the importance of tokenization, its role in LM performance and behavior is poorly understood due to the challenge of measuring the impact of tokenization in isolation. To address this need, we present TokSuite, a collection of models and a benchmark that supports research into tokenization’s influence on LMs. Specifically, we train fourteen models that use different tokenizers but are otherwise identical using the same architecture, dataset, training budget, and initialization. Additionally, we curate and release a new benchmark that specifically measures model performance subject to real-world perturbations that are likely to influence tokenization. Together, TokSuite allows robust decoupling of the influence of a model’s tokenizer, supporting a series of novel findings that elucidate the respective benefits and shortcomings of a wide range of popular tokenizers.

Method: Adversarial training framework with competitive dynamic between generator (user simulator) and discriminator, iteratively improving simulator realism through adversarial iterations.

Result: Fine-tuned simulators dramatically outperform zero-shot base models at surfacing system issues, with adversarial training enhancing diversity, distributional alignment, and predictive validity. The simulator achieves strong correlation between simulated and real failure rates while maintaining low distributional divergence of failure modes.

Conclusion: Adversarial training is a promising approach for creating realistic user simulators in mental health support TOD domains, enabling rapid, reliable, and cost-effective system evaluation before deployment.

Abstract: Realistic user simulation is crucial for training and evaluating task-oriented dialogue (TOD) systems, yet creating simulators that accurately replicate human behavior remains challenging. A key property of effective simulators is their ability to expose failure modes of the systems they evaluate. We present an adversarial training framework that iteratively improves user simulator realism through a competitive dynamic between a generator (user simulator) and a discriminator. Applied to mental health support chatbots, our approach demonstrates that fine-tuned simulators dramatically outperform zero-shot base models at surfacing system issues, and adversarial training further enhances diversity, distributional alignment, and predictive validity. The resulting simulator achieves a strong correlation between simulated and real failure occurrence rates across diverse chatbot configurations while maintaining low distributional divergence of failure modes. Discriminator accuracy decreases drastically after three adversarial iterations, suggesting improved realism. These results provide evidence that adversarial training is a promising approach for creating realistic user simulators in mental health support TOD domains, enabling rapid, reliable, and cost-effective system evaluation before deployment.

Method: Controlled turn-level comparison where expert human tutors, novice human tutors, and multiple LLMs respond to the same set of math remediation conversation turns. Analysis of instructional strategies (restating/revoicing, pressing for accuracy) and linguistic characteristics (lexical diversity, readability, politeness, agency).

Result: LLMs approach expert levels of perceived pedagogical quality on average but show systematic differences: they underuse restating/revoicing strategies characteristic of expert tutors, while producing longer, more lexically diverse, and more polite responses. Restating/revoicing, lexical diversity, and pressing for accuracy positively correlate with pedagogical quality, while higher agentic and polite language negatively correlate.

Conclusion: Recent LLMs exhibit pedagogical quality comparable to expert human tutors but rely on different instructional and linguistic strategies. Findings highlight the importance of analyzing both instructional strategies and linguistic characteristics when evaluating tutoring responses across human tutors and intelligent tutoring systems.

Abstract: Recent work has explored the use of large language models for generating tutoring responses in mathematics, yet it remains unclear how closely their instructional behavior aligns with expert human practice. We examine this question using a controlled, turn-level comparison in which expert human tutors, novice human tutors, and multiple large language models respond to the same set of math remediation conversation turns. We examine both instructional strategies and linguistic characteristics of tutoring responses, including restating and revoicing, pressing for accuracy, lexical diversity, readability, politeness, and agency. We find that large language models approach expert levels of perceived pedagogical quality on average but exhibit systematic differences in their instructional and linguistic profiles. In particular, large language models tend to underuse restating and revoicing strategies characteristic of expert human tutors, while producing longer, more lexically diverse, and more polite responses. Statistical analyses show that restating and revoicing, lexical diversity, and pressing for accuracy are positively associated with perceived pedagogical quality, whereas higher levels of agentic and polite language are negatively associated. Overall, recent large language models exhibit levels of perceived pedagogical quality comparable to expert human tutors, while relying on different instructional and linguistic strategies. These findings underscore the value of analyzing instructional strategies and linguistic characteristics when evaluating tutoring responses across human tutors and intelligent tutoring systems.

Method: The authors explore model editing algorithms (ROME, IKE, WISE) and design new editing targets specifically for an unlearning setting, adapting these editing approaches to the task of information removal rather than redirection.

Result: Model editing approaches can exceed baseline unlearning methods in terms of quality of forgetting, depending on the specific setting. However, they face similar challenges to traditional unlearning techniques.

Conclusion: While model editing algorithms show promise for unlearning tasks and can outperform existing unlearning methods, they still struggle with precisely defining what should be unlearned and maintaining overall model performance without damage.

Abstract: Machine unlearning aims to remove unwanted information from a model, but many methods are inefficient for LLMs with large numbers of parameters or fail to fully remove the intended information without degrading performance on knowledge that should be retained. Model editing algorithms solve a similar problem of changing information in models, but they focus on redirecting inputs to a new target rather than removing that information altogether. In this work, we explore the editing algorithms ROME, IKE, and WISE and design new editing targets for an unlearning setting. Through this investigation, we show that model editing approaches can exceed baseline unlearning methods in terms of quality of forgetting depending on the setting. Like traditional unlearning techniques, they struggle to encapsulate the scope of what is to be unlearned without damage to the overall model performance.

Method: Proposed FEND framework integrates speech and text modalities, leveraging 13 multi-lingual datasets across 5 languages, systematically evaluating multi-modal fusion performance.

Result: Multi-modal fusion excels in AD and depression detection but underperforms in ASD due to dataset heterogeneity. Modality imbalance is common, and cross-corpus experiments show robust performance in consistent scenarios but degradation in multi-lingual/task-heterogeneous settings.

Conclusion: FEND advances automated, lifespan-inclusive neuropsychiatric assessment by providing benchmarks and analysis of performance factors, encouraging adoption for fair comparisons and reproducible research.

Abstract: Neuropsychiatric disorders, such as Alzheimer’s disease (AD), depression, and autism spectrum disorder (ASD), are characterized by linguistic and acoustic abnormalities, offering potential biomarkers for early detection. Despite the promise of multi-modal approaches, challenges like multi-lingual generalization and the absence of a unified evaluation framework persist. To address these gaps, we propose FEND (Foundation model-based Evaluation of Neuropsychiatric Disorders), a comprehensive multi-modal framework integrating speech and text modalities for detecting AD, depression, and ASD across the lifespan. Leveraging 13 multi-lingual datasets spanning English, Chinese, Greek, French, and Dutch, we systematically evaluate multi-modal fusion performance. Our results show that multi-modal fusion excels in AD and depression detection but underperforms in ASD due to dataset heterogeneity. We also identify modality imbalance as a prevalent issue, where multi-modal fusion fails to surpass the best mono-modal models. Cross-corpus experiments reveal robust performance in task- and language-consistent scenarios but noticeable degradation in multi-lingual and task-heterogeneous settings. By providing extensive benchmarks and a detailed analysis of performance-influencing factors, FEND advances the field of automated, lifespan-inclusive, and multi-lingual neuropsychiatric disorder assessment. We encourage researchers to adopt the FEND framework for fair comparisons and reproducible research.

Method: Multi-task evaluation setup associating demographics with names, professions, and education levels; comparing attribution-based and correlation-based methods for locating bias features; using targeted sparse autoencoder feature ablations in Gemma-2-9B.

Result: Attribution-based ablations effectively reduce race and gender profession stereotypes while preserving name recognition accuracy, while correlation-based ablations work better for education bias. Removing attribution features in education tasks causes “prior collapse” and increases bias.

Conclusion: Demographic bias arises from task-specific mechanisms rather than absolute demographic markers, and mechanistic inference-time interventions enable surgical debiasing without compromising core model capabilities.

Abstract: We investigate how independent demographic bias mechanisms are from general demographic recognition in language models. Using a multi-task evaluation setup where demographics are associated with names, professions, and education levels, we measure whether models can be debiased while preserving demographic detection capabilities. We compare attribution-based and correlation-based methods for locating bias features. We find that targeted sparse autoencoder feature ablations in Gemma-2-9B reduce bias without degrading recognition performance: attribution-based ablations mitigate race and gender profession stereotypes while preserving name recognition accuracy, whereas correlation-based ablations are more effective for education bias. Qualitative analysis further reveals that removing attribution features in education tasks induces ``prior collapse’’, thus increasing overall bias. This highlights the need for dimension-specific interventions. Overall, our results show that demographic bias arises from task-specific mechanisms rather than absolute demographic markers, and that mechanistic inference-time interventions can enable surgical debiasing without compromising core model capabilities.

Method: Introduced semantic deceptions using novel symbols for digits and mathematical operators, then tested LLMs on simple calculations with this altered notation to assess abstraction capacity and resistance to misleading semantic cues.

Result: Semantic cues significantly deteriorate LLM performance on simple tasks, revealing limitations in symbolic manipulation and a tendency to over-rely on surface-level semantics, with chain-of-thought potentially amplifying statistical correlation reliance.

Conclusion: LLMs have fundamental limitations in symbolic reasoning, undermining claims of genuine reasoning abilities and raising ethical concerns for decision-making contexts where robust abstraction is essential.

Abstract: Large language models (LLMs) are increasingly used in situations where human values are at stake, such as decision-making tasks that involve reasoning when performed by humans. We investigate the so-called reasoning capabilities of LLMs over novel symbolic representations by introducing an experimental framework that tests their ability to process and manipulate unfamiliar symbols. We introduce semantic deceptions: situations in which symbols carry misleading semantic associations due to their form, such as being embedded in specific contexts, designed to probe whether LLMs can maintain symbolic abstraction or whether they default to exploiting learned semantic associations. We redefine standard digits and mathematical operators using novel symbols, and task LLMs with solving simple calculations expressed in this altered notation. The objective is: (1) to assess LLMs’ capacity for abstraction and manipulation of arbitrary symbol systems; (2) to evaluate their ability to resist misleading semantic cues that conflict with the task’s symbolic logic. Through experiments with four LLMs we show that semantic cues can significantly deteriorate reasoning models’ performance on very simple tasks. They reveal limitations in current LLMs’ ability for symbolic manipulations and highlight a tendency to over-rely on surface-level semantics, suggesting that chain-of-thoughts may amplify reliance on statistical correlations. Even in situations where LLMs seem to correctly follow instructions, semantic cues still impact basic capabilities. These limitations raise ethical and societal concerns, undermining the widespread and pernicious tendency to attribute reasoning abilities to LLMs and suggesting how LLMs might fail, in particular in decision-making contexts where robust symbolic reasoning is essential and should not be compromised by residual semantic associations inherited from the model’s training.

Method: EssayCBM uses a rubric-aligned framework with eight dedicated prediction heads on an encoder to evaluate specific writing concepts (Thesis Clarity, Evidence Use, etc.). Concept scores form a transparent bottleneck, and a lightweight network computes the final grade using only these concept scores. The system allows instructors to adjust concept predictions and instantly see updated grades.

Result: EssayCBM matches the performance of black-box grading systems while providing interpretable, concept-level feedback. The framework offers an intuitive web interface for human-in-the-loop evaluation where instructors can adjust concept scores and see immediate grade updates.

Conclusion: EssayCBM successfully addresses the interpretability problem in automated essay grading by providing transparent, concept-based assessment that maintains performance parity with black-box models while enabling accountable human-in-the-loop evaluation through an intuitive interface.

Abstract: Understanding how automated grading systems evaluate essays remains a significant challenge for educators and students, especially when large language models function as black boxes. We introduce EssayCBM, a rubric-aligned framework that prioritizes interpretability in essay assessment. Instead of predicting grades directly from text, EssayCBM evaluates eight writing concepts, such as Thesis Clarity and Evidence Use, through dedicated prediction heads on an encoder. These concept scores form a transparent bottleneck, and a lightweight network computes the final grade using only concepts. Instructors can adjust concept predictions and instantly view the updated grade, enabling accountable human-in-the-loop evaluation. EssayCBM matches black-box performance while offering actionable, concept-level feedback through an intuitive web interface.

Method: Introduces MediEval benchmark linking MIMIC-IV EHRs to unified knowledge base (UMLS/biomedical vocabularies). Generates diverse factual/counterfactual statements within real patient contexts. Uses 4-quadrant framework considering knowledge grounding and contextual consistency. Proposes Counterfactual Risk-Aware Fine-tuning (CoRFu), a DPO-based method with asymmetric penalty targeting unsafe confusions.

Result: Identifies critical failure modes including hallucinated support and truth inversion that current LLMs frequently exhibit. CoRFu improves by +16.4 macro-F1 points over base model and eliminates truth inversion errors, demonstrating higher accuracy and substantially greater safety.

Conclusion: MediEval provides systematic evaluation framework for medical LLMs, revealing safety risks. CoRFu effectively addresses these risks through targeted fine-tuning, enabling safer adoption of LLMs in medicine.

Abstract: Large Language Models (LLMs) are increasingly applied to medicine, yet their adoption is limited by concerns over reliability and safety. Existing evaluations either test factual medical knowledge in isolation or assess patient-level reasoning without verifying correctness, leaving a critical gap. We introduce MediEval, a benchmark that links MIMIC-IV electronic health records (EHRs) to a unified knowledge base built from UMLS and other biomedical vocabularies. MediEval generates diverse factual and counterfactual medical statements within real patient contexts, enabling systematic evaluation across a 4-quadrant framework that jointly considers knowledge grounding and contextual consistency. Using this framework, we identify critical failure modes, including hallucinated support and truth inversion, that current proprietary, open-source, and domain-specific LLMs frequently exhibit. To address these risks, we propose Counterfactual Risk-Aware Fine-tuning (CoRFu), a DPO-based method with an asymmetric penalty targeting unsafe confusions. CoRFu improves by +16.4 macro-F1 points over the base model and eliminates truth inversion errors, demonstrating both higher accuracy and substantially greater safety.

Method: Hybrid Mamba-Transformer architecture with Mixture-of-Experts, pretrained on 25 trillion tokens (including 3+ trillion new tokens over Nemotron 2), followed by supervised fine-tuning and large-scale reinforcement learning on diverse environments.

Result: Achieves better accuracy than Nemotron 2 Nano while activating less than half parameters per forward pass, 3.3x higher inference throughput than similar-sized open models, enhanced agentic/reasoning/chat abilities, and 1M token context support.

Conclusion: Nemotron 3 Nano demonstrates superior efficiency and performance compared to previous generation and similar-sized models, with released checkpoints available on Hugging Face for both base and post-trained versions.

Abstract: We present Nemotron 3 Nano 30B-A3B, a Mixture-of-Experts hybrid Mamba-Transformer language model. Nemotron 3 Nano was pretrained on 25 trillion text tokens, including more than 3 trillion new unique tokens over Nemotron 2, followed by supervised fine tuning and large-scale RL on diverse environments. Nemotron 3 Nano achieves better accuracy than our previous generation Nemotron 2 Nano while activating less than half of the parameters per forward pass. It achieves up to 3.3x higher inference throughput than similarly-sized open models like GPT-OSS-20B and Qwen3-30B-A3B-Thinking-2507, while also being more accurate on popular benchmarks. Nemotron 3 Nano demonstrates enhanced agentic, reasoning, and chat abilities and supports context lengths up to 1M tokens. We release both our pretrained Nemotron 3 Nano 30B-A3B Base and post-trained Nemotron 3 Nano 30B-A3B checkpoints on Hugging Face.

Method: Evaluate established strategies for handling unknown total relevant documents by measuring correlation between retrieval quality metrics and LLM-based judgments of response quality. Experiments across multiple datasets with 2-15 relevant documents. Introduce a new simple retrieval quality measure that performs well without requiring knowledge of total relevant documents.

Result: Experimental results show performance of different metrics in correlating with LLM-based response quality judgments. The newly introduced simple measure performs well without needing to know total number of relevant documents.

Conclusion: Proposed retrieval quality measure provides effective evaluation in realistic settings where total relevant documents are unknown, addressing a practical limitation in retrieval system evaluation.

Abstract: In realistic retrieval settings with large and evolving knowledge bases, the total number of documents relevant to a query is typically unknown, and recall cannot be computed. In this paper, we evaluate several established strategies for handling this limitation by measuring the correlation between retrieval quality metrics and LLM-based judgments of response quality, where responses are generated from the retrieved documents. We conduct experiments across multiple datasets with a relatively low number of relevant documents (2-15). We also introduce a simple retrieval quality measure that performs well without requiring knowledge of the total number of relevant documents.

Method: Created “Hearing to Translate” - a comprehensive test suite benchmarking 5 state-of-the-art SpeechLLMs against 16 strong direct and cascade systems. Evaluation spanned 16 benchmarks, 13 language pairs, and 9 challenging conditions including disfluent, noisy, and long-form speech.

Result: Cascaded systems remain the most reliable overall. Current SpeechLLMs only match cascades in selected settings, and speech foundation models (SFMs) lag behind both approaches. Integrating an LLM (either within the model or in a pipeline) is essential for high-quality speech translation.

Conclusion: While SpeechLLMs represent an interesting direction for speech translation, traditional cascade architectures combining speech foundation models with multilingual LLMs currently provide more reliable performance across diverse conditions.

Abstract: As Large Language Models (LLMs) expand beyond text, integrating speech as a native modality has given rise to SpeechLLMs, which aim to translate spoken language directly, thereby bypassing traditional transcription-based pipelines. Whether this integration improves speech-to-text translation quality over established cascaded architectures, however, remains an open question. We present Hearing to Translate, the first comprehensive test suite rigorously benchmarking 5 state-of-the-art SpeechLLMs against 16 strong direct and cascade systems that couple leading speech foundation models (SFM), with multilingual LLMs. Our analysis spans 16 benchmarks, 13 language pairs, and 9 challenging conditions, including disfluent, noisy, and long-form speech. Across this extensive evaluation, we find that cascaded systems remain the most reliable overall, while current SpeechLLMs only match cascades in selected settings and SFMs lag behind both, highlighting that integrating an LLM, either within the model or in a pipeline, is essential for high-quality speech translation.

Method: Uses Mixture-of-Experts hybrid Mamba-Transformer architecture with LatentMoE (for Super/Ultra models) and MTP layers for faster text generation. All models are post-trained using multi-environment reinforcement learning for reasoning and tool use capabilities.

Result: Nano outperforms comparable models in accuracy while being cost-efficient; Super is optimized for collaborative agents and IT automation; Ultra provides state-of-the-art accuracy and reasoning performance. All models support up to 1M token context.

Conclusion: Nemotron 3 family offers scalable solutions from cost-efficient Nano to state-of-the-art Ultra, with open release of weights, software, and recipes, enabling broad adoption across different application domains.

Abstract: We introduce the Nemotron 3 family of models - Nano, Super, and Ultra. These models deliver strong agentic, reasoning, and conversational capabilities. The Nemotron 3 family uses a Mixture-of-Experts hybrid Mamba-Transformer architecture to provide best-in-class throughput and context lengths of up to 1M tokens. Super and Ultra models are trained with NVFP4 and incorporate LatentMoE, a novel approach that improves model quality. The two larger models also include MTP layers for faster text generation. All Nemotron 3 models are post-trained using multi-environment reinforcement learning enabling reasoning, multi-step tool use, and support granular reasoning budget control. Nano, the smallest model, outperforms comparable models in accuracy while remaining extremely cost-efficient for inference. Super is optimized for collaborative agents and high-volume workloads such as IT ticket automation. Ultra, the largest model, provides state-of-the-art accuracy and reasoning performance. Nano is released together with its technical report and this white paper, while Super and Ultra will follow in the coming months. We will openly release the model weights, pre- and post-training software, recipes, and all data for which we hold redistribution rights.

Method: Progressive introduction of nonlinearities, self-attention, and multi-layer transformers; evaluation on character-level Tiny Shakespeare and word-level PTB/WikiText-2; comparison using test NLL, parameter count, and training FLOPs.

Result: Attention-based models dominate MLPs in per-FLOP efficiency even at small scale; increasing depth/context without sufficient optimization can degrade performance; RoPE and other large-model techniques don’t necessarily transfer to small-model regimes.

Conclusion: Small language models have different optimization dynamics than large ones, with attention being efficient even at small scale, but careful architectural choices are needed as techniques from large models don’t always scale down effectively.

Abstract: We present a systematic empirical study of small language models under strict compute constraints, analyzing how architectural choices and training budget interact to determine performance. Starting from a linear next-token predictor, we progressively introduce nonlinearities, self-attention, and multi-layer transformer architectures, evaluating each on character-level modeling of Tiny Shakespeare and word-level modeling of Penn Treebank (PTB) and WikiText-2. We compare models using test negative log-likelihood (NLL), parameter count, and approximate training FLOPs to characterize accuracy-efficiency trade-offs. Our results show that attention-based models dominate MLPs in per-FLOP efficiency even at small scale, while increasing depth or context without sufficient optimization can degrade performance. We further examine rotary positional embeddings (RoPE), finding that architectural techniques successful in large language models do not necessarily transfer to small-model regimes.

Method: Cross-model Reasoning Distillation Provenance Tracing framework that compares predictive probabilities from teacher, original student, and distilled models to classify each action’s origin, plus a teacher-guided data selection method based on teacher-student divergence.

Result: The framework shows distilled models can generate teacher-originated actions in test-time contexts, which correlate with performance, and the data selection method proves effective across multiple teacher-student model combinations.

Conclusion: The provenance-tracing framework provides valuable insights into reasoning distillation and shows promise for improving distillation techniques through principled data selection.

Abstract: Reasoning distillation has attracted increasing attention. It typically leverages a large teacher model to generate reasoning paths, which are then used to fine-tune a student model so that it mimics the teacher’s behavior in training contexts. However, previous approaches have lacked a detailed analysis of the origins of the distilled model’s capabilities. It remains unclear whether the student can maintain consistent behaviors with the teacher in novel test-time contexts, or whether it regresses to its original output patterns, raising concerns about the generalization of distillation models. To analyse this question, we introduce a cross-model Reasoning Distillation Provenance Tracing framework. For each action (e.g., a sentence) produced by the distilled model, we obtain the predictive probabilities assigned by the teacher, the original student, and the distilled model under the same context. By comparing these probabilities, we classify each action into different categories. By systematically disentangling the provenance of each action, we experimentally demonstrate that, in test-time contexts, the distilled model can indeed generate teacher-originated actions, which correlate with and plausibly explain observed performance on distilled model. Building on this analysis, we further propose a teacher-guided data selection method. Unlike prior approach that rely on heuristics, our method directly compares teacher-student divergences on the training data, providing a principled selection criterion. We validate the effectiveness of our approach across multiple representative teacher models and diverse student models. The results highlight the utility of our provenance-tracing framework and underscore its promise for reasoning distillation. We hope to share Reasoning Distillation Provenance Tracing and our insights into reasoning distillation with the community.

Method: Neural network framework for token-level hallucination detection: 1) Freeze LLM parameters, 2) Use lightweight MLP probes for nonlinear modeling of high-level hidden states, 3) Multi-objective joint loss function for detection stability and semantic disambiguity, 4) Bayesian optimization to automatically search for optimal probe insertion layers via layer position-probe performance response model.

Result: Experiments on LongFact, HealthBench, and TriviaQA show MLP probes significantly outperform state-of-the-art methods in accuracy, recall, and detection capability under low false-positive conditions.

Conclusion: MLP probes provide effective real-time hallucination detection by capturing nonlinear semantic structures in LLM hidden states, overcoming limitations of linear probes and uncertainty/retrieval-based methods, enabling safer LLM applications in high-risk domains.

Abstract: Large language models(LLMs) excel at text generation and knowledge question-answering tasks, but they are prone to generating hallucinated content, severely limiting their application in high-risk domains. Current hallucination detection methods based on uncertainty estimation and external knowledge retrieval suffer from the limitation that they still produce erroneous content at high confidence levels and rely heavily on retrieval efficiency and knowledge coverage. In contrast, probe methods that leverage the model’s hidden-layer states offer real-time and lightweight advantages. However, traditional linear probes struggle to capture nonlinear structures in deep semantic spaces.To overcome these limitations, we propose a neural network-based framework for token-level hallucination detection. By freezing language model parameters, we employ lightweight MLP probes to perform nonlinear modeling of high-level hidden states. A multi-objective joint loss function is designed to enhance detection stability and semantic disambiguity. Additionally, we establish a layer position-probe performance response model, using Bayesian optimization to automatically search for optimal probe insertion layers and achieve superior training results.Experimental results on LongFact, HealthBench, and TriviaQA demonstrate that MLP probes significantly outperform state-of-the-art methods in accuracy, recall, and detection capability under low false-positive conditions.

Method: TriAligner uses dual-encoder architecture with contrastive learning, incorporates both native and English translations across modalities, employs data preprocessing/augmentation with LLMs, and uses hard negative sampling for representation learning.

Result: Significant improvements in retrieval accuracy and fact-checking performance on both monolingual and crosslingual benchmarks compared to baselines.

Conclusion: The proposed TriAligner system effectively addresses multilingual fact-checked claim retrieval by learning relative importance of different sources and enhancing robustness through various techniques.

Abstract: This paper presents our system for SemEval-2025 Task 7: Multilingual and Crosslingual Fact-Checked Claim Retrieval. In an era where misinformation spreads rapidly, effective fact-checking is increasingly critical. We introduce TriAligner, a novel approach that leverages a dual-encoder architecture with contrastive learning and incorporates both native and English translations across different modalities. Our method effectively retrieves claims across multiple languages by learning the relative importance of different sources in alignment. To enhance robustness, we employ efficient data preprocessing and augmentation using large language models while incorporating hard negative sampling to improve representation learning. We evaluate our approach on monolingual and crosslingual benchmarks, demonstrating significant improvements in retrieval accuracy and fact-checking performance over baselines.

Method: Proposed reflection pretraining that enables biological sequence models to generate auxiliary “thinking tokens” beyond standard answer tokens, theoretically enhancing language expressiveness and enabling intermediate reasoning.

Result: The augmented token set significantly enhances biological language expressiveness, and the pretraining approach teaches protein models to self-correct, leading to substantial performance gains compared to standard pretraining.

Conclusion: Reflection pretraining enables Chain-of-Thought reasoning in biological sequence models for the first time, overcoming token space limitations and improving reasoning capacity through intermediate “thinking tokens”.

Abstract: Chain-of-Thought (CoT) prompting has significantly advanced task-solving capabilities in natural language processing with large language models. Unlike standard prompting, CoT encourages the model to generate intermediate reasoning steps, non-answer tokens, that help guide the model toward more accurate final outputs. These intermediate steps enable more complex reasoning processes such as error correction, memory management, future planning, and self-reflection. However, applying CoT to non-natural language domains, such as protein and RNA language models, is not yet possible, primarily due to the limited expressiveness of their token spaces (e.g., amino acid tokens). In this work, we propose and define the concept of language expressiveness: the ability of a given language, using its tokens and grammar, to encode information. We show that the limited expressiveness of protein language severely restricts the applicability of CoT-style reasoning. To overcome this, we introduce reflection pretraining, for the first time in a biological sequence model, which enables the model to engage in intermediate reasoning through the generation of auxiliary “thinking tokens” beyond simple answer tokens. Theoretically, we demonstrate that our augmented token set significantly enhances biological language expressiveness, thereby improving the overall reasoning capacity of the model. Experimentally, our pretraining approach teaches protein models to self-correct and leads to substantial performance gains compared to standard pretraining.

Method: Three-step pipeline: (1) creates complex conversational data between LLM patients and LLM doctors, (2) evaluates conversations with a committee of 2 LLM judges across multiple dimensions, (3) converts identified mistakes into simplified single-shot QA scenarios.

Result: Created MedMistake-All (3,390 QA pairs where GPT-5 and Gemini 2.5 Pro fail) and MedMistake-Bench (211 expert-validated questions). Evaluation of 12 frontier LLMs showed GPT models, Claude, and Grok performed best on the validated benchmark.

Conclusion: MedMistake provides an automated pipeline for extracting and benchmarking clinical reasoning mistakes, offering valuable datasets for evaluating LLM performance in medical contexts and identifying areas where current models struggle.

Abstract: Large language models (LLMs) are increasingly evaluated in clinical settings using multi-dimensional rubrics which quantify reasoning quality, safety, and patient-centeredness. Yet, replicating specific mistakes in other LLM models is not straightforward and often requires manual effort. We introduce MedMistake, an automatic pipeline that extracts mistakes LLMs make in patient-doctor conversations and converts them into a benchmark of single-shot QA pairs. Our pipeline (1) creates complex, conversational data between an LLM patient and LLM doctor, (2) runs an evaluation with a committee of 2 LLM judges across a variety of dimensions and (3) creates simplified single-shot QA scenarios from those mistakes. We release MedMistake-All, a dataset of 3,390 single-shot QA pairs where GPT-5 and Gemini 2.5 Pro are currently failing to answer correctly, as judged by two LLM judges. We used medical experts to validate a subset of 211/3390 questions (MedMistake-Bench), which we used to run a final evaluation of 12 frontier LLMs: Claude Opus 4.5, Claude Sonnet 4.5, DeepSeek-Chat, Gemini 2.5 Pro, Gemini 3 Pro, GPT-4o, GPT-5, GPT-5.1, GPT-5.2, Grok 4, Grok 4.1, Mistral Large. We found that GPT models, Claude and Grok obtained the best performance on MedMistake-Bench. We release both the doctor-validated benchmark (MedMistake-Bench), as well as the full dataset (MedMistake-All) at https://huggingface.co/datasets/TheLumos/MedicalMistakeBenchmark.

Method: Analyze supervision allocation across prompt, chain-of-thought, and answer segments. Develop selective knowledge distillation focusing only on CoT tokens, and establish a truncation protocol to quantify computation-quality tradeoffs based on sequence length.

Result: Training on only the first 50% of tokens retains ≈94% of full-sequence performance on math benchmarks while reducing training time, memory usage, and FLOPs by about 50% each. CoT tokens alone can be effective when they encompass prompt and answer information.

Conclusion: Reasoning distillation benefits from prioritizing early reasoning tokens, and selective supervision on CoT segments provides a simple lever for computation-quality tradeoffs, enabling more efficient model distillation.

Abstract: Distilling the reasoning capabilities from a large language model (LLM) to a smaller student model often involves training on substantial amounts of reasoning data. However, distillation over lengthy sequences with prompt (P), chain-of-thought (CoT), and answer (A) segments makes the process computationally expensive. In this work, we investigate how the allocation of supervision across different segments (P, CoT, A) affects student performance. Our analysis shows that selective knowledge distillation over only the CoT tokens can be effective when the prompt and answer information is encompassed by it. Building on this insight, we establish a truncation protocol to quantify computation-quality tradeoffs as a function of sequence length. We observe that training on only the first $50%$ of tokens of every training sequence can retain, on average, $\approx94%$ of full-sequence performance on math benchmarks while reducing training time, memory usage, and FLOPs by about $50%$ each. These findings suggest that reasoning distillation benefits from prioritizing early reasoning tokens and provides a simple lever for computation-quality tradeoffs. Codes are available at https://github.com/weiruichen01/distilling-the-essence.

Method: SFTKey uses a two-stage training scheme: 1) conventional SFT to ensure proper output format, 2) fine-tuning only on the Key portion (final answer) to improve accuracy while maintaining format correctness.

Result: Extensive experiments across multiple benchmarks and model families show SFTKey achieves an average accuracy improvement exceeding 5% over conventional SFT while preserving the ability to generate correct formats.

Conclusion: This study advances LLM fine-tuning by explicitly balancing CoT learning with additional optimization on answer-relevant tokens, addressing attention imbalance in conventional SFT approaches.

Abstract: With the rapid advancement of Large Language Models (LLMs), the Chain-of-Thought (CoT) component has become significant for complex reasoning tasks. However, in conventional Supervised Fine-Tuning (SFT), the model could allocate disproportionately more attention to CoT sequences with excessive length. This reduces focus on the much shorter but essential Key portion-the final answer, whose correctness directly determines task success and evaluation quality. To address this limitation, we propose SFTKey, a two-stage training scheme. In the first stage, conventional SFT is applied to ensure proper output format, while in the second stage, only the Key portion is fine-tuned to improve accuracy. Extensive experiments across multiple benchmarks and model families demonstrate that SFTKey achieves an average accuracy improvement exceeding 5% over conventional SFT, while preserving the ability to generate correct formats. Overall, this study advances LLM fine-tuning by explicitly balancing CoT learning with additional optimization on answer-relevant tokens.

Method: Data-Adaptive Semantic Refinement (DAS) framework couples fixed GNN with LLM in closed feedback loop. GNN provides implicit supervisory signals to guide LLM’s semantic refinement, and refined semantics are fed back to update the same graph learner. Works on both text-rich and text-free graphs.

Result: Consistent improvements on structure-dominated graphs while remaining competitive on semantics-rich graphs. Demonstrates effectiveness of data-centric semantic adaptation under structure-semantics heterogeneity.

Conclusion: Data-centric approach to semantic adaptation (rather than model-centric inductive bias injection) effectively addresses structure-semantics heterogeneity in graph learning, enabling better generalization across diverse graph domains.

Abstract: Graph-structured data exhibit substantial heterogeneity in where their predictive signals originate: in some domains, node-level semantics dominate, while in others, structural patterns play a central role. This structure-semantics heterogeneity implies that no graph learning model with a fixed inductive bias can generalize optimally across diverse graph domains. However, most existing methods address this challenge from the model side by incrementally injecting new inductive biases, which remains fundamentally limited given the open-ended diversity of real-world graphs. In this work, we take a data-centric perspective and treat node semantics as a task-adaptive variable. We propose a Data-Adaptive Semantic Refinement framework DAS for graph representation learning, which couples a fixed graph neural network (GNN) and a large language model (LLM) in a closed feedback loop. The GNN provides implicit supervisory signals to guide the semantic refinement of LLM, and the refined semantics are fed back to update the same graph learner. We evaluate our approach on both text-rich and text-free graphs. Results show consistent improvements on structure-dominated graphs while remaining competitive on semantics-rich graphs, demonstrating the effectiveness of data-centric semantic adaptation under structure-semantics heterogeneity.

Method: The authors propose using semi-supervised learning techniques that leverage both labeled and unlabeled data. They emphasize the importance of task-specific augmentations rather than general-purpose ones, applying these methods to both LLM prompts and responses for safety classification.

Result: The paper demonstrates that semi-supervised learning significantly improves safety task performance compared to supervised-only approaches. Task-specific augmentations are shown to be crucial, providing substantially better results than general-purpose augmentation techniques.

Conclusion: Semi-supervised learning with task-specific augmentations offers an effective solution to data scarcity problems in LLM safety classification, providing improved performance while reducing reliance on large labeled datasets.

Abstract: Safety for Large Language Models (LLMs) has been an ongoing research focus since their emergence and is even more relevant nowadays with the increasing capacity of those models. Currently, there are several guardrails in place for all public LLMs and multiple proposed datasets for training safety classifiers. However, training these safety classifiers relies on large quantities of labeled data, which can be problematic to acquire, prone to labeling errors, or often include synthetic data. To address these issues, we suggest a different approach: utilizing semi-supervised learning techniques, which leverage both labeled and unlabeled data, to improve the performance on the safety task. We analyze the improvements that these techniques can offer for both prompts given to Large Language Models and the responses to those requests. Moreover, since augmentation is the central part of semi-supervised algorithms, we demonstrate the importance of using task-specific augmentations, which significantly increase the performance when compared to general-purpose augmentation techniques.

Method: Created ClarifyMT-Bench using a hybrid LLM-human pipeline with a five-dimensional ambiguity taxonomy and six behaviorally diverse simulated user personas, generating 6,120 multi-turn dialogues. Proposed ClarifyAgent approach that decomposes clarification into perception, forecasting, tracking, and planning components.

Result: Evaluation of ten representative LLMs revealed consistent under-clarification bias - LLMs tend to answer prematurely, and performance degrades as dialogue depth increases. ClarifyAgent substantially improved robustness across ambiguity conditions.

Conclusion: ClarifyMT-Bench establishes a reproducible foundation for studying when LLMs should ask vs. answer questions and how to navigate ambiguity in real-world human-LLM interactions, with ClarifyAgent showing promise for improving clarification capabilities.

Abstract: Large language models (LLMs) are increasingly deployed as conversational assistants in open-domain, multi-turn settings, where users often provide incomplete or ambiguous information. However, existing LLM-focused clarification benchmarks primarily assume single-turn interactions or cooperative users, limiting their ability to evaluate clarification behavior in realistic settings. We introduce \textbf{ClarifyMT-Bench}, a benchmark for multi-turn clarification grounded in a five-dimensional ambiguity taxonomy and a set of six behaviorally diverse simulated user personas. Through a hybrid LLM-human pipeline, we construct 6,120 multi-turn dialogues capturing diverse ambiguity sources and interaction patterns. Evaluating ten representative LLMs uncovers a consistent under-clarification bias: LLMs tend to answer prematurely, and performance degrades as dialogue depth increases. To mitigate this, we propose \textbf{ClarifyAgent}, an agentic approach that decomposes clarification into perception, forecasting, tracking, and planning, substantially improving robustness across ambiguity conditions. ClarifyMT-Bench establishes a reproducible foundation for studying when LLMs should ask, when they should answer, and how to navigate ambiguity in real-world human-LLM interactions.

Method: Proposes SpidR-Adapt with multi-task adaptive pre-training (MAdaPT) as a bi-level optimization framework. Uses first-order bi-level optimization (FOBLO) for scalable meta-training and stabilizes training with interleaved supervision alternating self-supervised and supervised objectives.

Result: Achieves rapid gains in phonemic discriminability (ABX) and spoken language modeling (sWUGGY, sBLIMP, tSC), improving over in-domain language models after training on less than 1 hour of target-language audio - 100x more data-efficient than standard training.

Conclusion: Provides a practical, architecture-agnostic path toward biologically inspired, data-efficient speech representations, bridging the gap between human language acquisition efficiency and machine learning models.

Abstract: Human infants, with only a few hundred hours of speech exposure, acquire basic units of new languages, highlighting a striking efficiency gap compared to the data-hungry self-supervised speech models. To address this gap, this paper introduces SpidR-Adapt for rapid adaptation to new languages using minimal unlabeled data. We cast such low-resource speech representation learning as a meta-learning problem and construct a multi-task adaptive pre-training (MAdaPT) protocol which formulates the adaptation process as a bi-level optimization framework. To enable scalable meta-training under this framework, we propose a novel heuristic solution, first-order bi-level optimization (FOBLO), avoiding heavy computation costs. Finally, we stabilize meta-training by using a robust initialization through interleaved supervision which alternates self-supervised and supervised objectives. Empirically, SpidR-Adapt achieves rapid gains in phonemic discriminability (ABX) and spoken language modeling (sWUGGY, sBLIMP, tSC), improving over in-domain language models after training on less than 1h of target-language audio, over $100\times$ more data-efficient than standard training. These findings highlight a practical, architecture-agnostic path toward biologically inspired, data-efficient representations. We open-source the training code and model checkpoints at https://github.com/facebookresearch/spidr-adapt.

Method: SMART uses hierarchical processing with three components: 1) Grammarian Tree LSTM for syntax-aware fact extraction (subject-relation-object), 2) Compact indexed MANN memory storing facts as 384D vectors with source tracking, 3) 6-layer Transformer for fusing retrieved facts into responses.

Result: SMART uses only 45.51M parameters (64% less than GPT-2, 69% less than BERT) and achieves 21.3% higher accuracy than GPT-2. It supports dual inference modes: fast path for known documents (sub-second) and dynamic path with RAG for new uploads.

Conclusion: SMART provides a practical solution for engineering manual comprehension with better accuracy, reduced hallucinations, and lower computational requirements than comparable small transformer models through structured memory and reasoning.

Abstract: The user of Engineering Manuals (EM) finds it difficult to read EM s because they are long, have a dense format which includes written documents, step by step procedures, and standard parameter lists for engineering equipment. Off the shelf transformers, especially compact ones, treat this material as a flat stream of tokens. This approach leads to confident but incorrect numeric answers and forces the models to memorize separate facts inefficiently. SMART (Structured Memory and Reasoning Transformer) offers a different and practical solution to the above problem. SMART structures its processing by using a hierarchical approach, and is based upon three main job categories (1) A syntax-aware Fact Extractor (Grammarian) Tree LSTM which extracts facts as subject relation object relations from EM sentences (2) A compact indexed memory MANN (Memory Augmented Neural Network) that indexes these Rational Subject Relation Objects as 384 dimensional vectors that are associated with the source of the information, and (3) A 6 layer Transformer that learns to fuse the previously retrieved facts into its generated response. The entire SMART model utilizes 45.51M parameters, which is 64% less than GPT-2 (124M) and 69% less than BERT (133M), and it achieves a 21.3% higher accuracy than GPT-2, indicating that SMART fits the data better with the least amount of processing requirements. SMART employs dual modes of inference an indexed fast path for known documents (sub-second answer times) and an indexed dynamic path assisted by RAGs for new uploads (FAISS Top 20 results with memory severed at 64 slots). In real world deployment, this framework leads to more well supported results with reduced hallucinations than comparable small transformer models.

Method: Jointly predicts multiple dependent tokens in a single transformer call by incorporating sampling procedure into the model. Trained via distillation or inverse autoregressive training without a teacher.

Result: Achieves state-of-the-art speculative decoding performance on Vicuna-7B, accepting over four tokens per step on Spec-Bench. Proves PTP can represent arbitrary autoregressive sequence distributions.

Conclusion: Parallel generation of long sequences is feasible without loss of modeling power, indicating the universality of the PTP framework.

Abstract: We propose Parallel Token Prediction (PTP), a universal framework for parallel sequence generation in language models. PTP jointly predicts multiple dependent tokens in a single transformer call by incorporating the sampling procedure into the model. This reduces the latency bottleneck of autoregressive decoding, and avoids the restrictive independence assumptions common in existing multi-token prediction methods. We prove that PTP can represent arbitrary autoregressive sequence distributions. PTP is trained either by distilling an existing model or through inverse autoregressive training without a teacher. Experimentally, we achieve state-of-the-art speculative decoding performance on Vicuna-7B by accepting over four tokens per step on Spec-Bench. The universality of our framework indicates that parallel generation of long sequences is feasible without loss of modeling power.

Method: Introduce a two-stage experimental pipeline that separates perception and reasoning: (1) perception stage converts each image independently to natural-language descriptions, (2) reasoning stage induces and applies rules using these descriptions, preventing cross-image inductive signal leakage.

Result: Across Mini-ARC, ACRE, and Bongard-LOGO datasets, perception capability is the dominant factor in performance gaps; manual inspection shows ~80% of VLM failures stem from perception errors, not reasoning deficiencies.

Conclusion: ARC-style benchmarks conflate perceptual and reasoning challenges, overstating machine reasoning deficiencies; evaluation protocols should disentangle perception from reasoning when assessing machine intelligence progress.

Abstract: Reasoning benchmarks such as the Abstraction and Reasoning Corpus (ARC) and ARC-AGI are widely used to assess progress in artificial intelligence and are often interpreted as probes of core, so-called ``fluid’’ reasoning abilities. Despite their apparent simplicity for humans, these tasks remain challenging for frontier vision-language models (VLMs), a gap commonly attributed to deficiencies in machine reasoning. We challenge this interpretation and hypothesize that the gap arises primarily from limitations in visual perception rather than from shortcomings in inductive reasoning. To verify this hypothesis, we introduce a two-stage experimental pipeline that explicitly separates perception and reasoning. In the perception stage, each image is independently converted into a natural-language description, while in the reasoning stage a model induces and applies rules using these descriptions. This design prevents leakage of cross-image inductive signals and isolates reasoning from perception bottlenecks. Across three ARC-style datasets, Mini-ARC, ACRE, and Bongard-LOGO, we show that the perception capability is the dominant factor underlying the observed performance gap by comparing the two-stage pipeline with against standard end-to-end one-stage evaluation. Manual inspection of reasoning traces in the VLM outputs further reveals that approximately 80 percent of model failures stem from perception errors. Together, these results demonstrate that ARC-style benchmarks conflate perceptual and reasoning challenges and that observed performance gaps may overstate deficiencies in machine reasoning. Our findings underscore the need for evaluation protocols that disentangle perception from reasoning when assessing progress in machine intelligence.

Method: Built on Qwen-2.5-Coder backbones with a Pooling by Multihead Attention (PMA) module that generates sequence embeddings from token embeddings, trained on 3 million publicly available data points.

Result: C2LLM models set new records on MTEB-Code among similar-sized models, with C2LLM-7B ranking 1st on the overall leaderboard.

Conclusion: C2LLM demonstrates that PMA-based sequence embedding generation effectively addresses limitations of traditional EOS-based approaches while achieving state-of-the-art performance in code embedding tasks.

Abstract: We present C2LLM - Contrastive Code Large Language Models, a family of code embedding models in both 0.5B and 7B sizes. Building upon Qwen-2.5-Coder backbones, C2LLM adopts a Pooling by Multihead Attention (PMA) module for generating sequence embedding from token embeddings, effectively 1) utilizing the LLM’s causal representations acquired during pretraining, while also 2) being able to aggregate information from all tokens in the sequence, breaking the information bottleneck in EOS-based sequence embeddings, and 3) supporting flexible adaptation of embedding dimension, serving as an alternative to MRL. Trained on three million publicly available data, C2LLM models set new records on MTEB-Code among models of similar sizes, with C2LLM-7B ranking 1st on the overall leaderboard.

Method: Introduced Denoising Entropy as a computable metric to quantify cumulative predictive uncertainty along generative paths. Proposed two algorithms: post-hoc selection method and real-time guidance strategy to optimize decoding paths.

Result: Entropy-guided methods significantly improve generation quality, consistently boosting accuracy on challenging reasoning, planning, and code benchmarks.

Conclusion: Denoising Entropy serves as a principled tool for understanding and controlling generation in MDMs, turning uncertainty from a liability into an advantage for discovering high-quality solutions.

Abstract: Masked Diffusion Models (MDMs) offer flexible, non-autoregressive generation, but this freedom introduces a challenge: final output quality is highly sensitive to the decoding order. We are the first to formalize this issue, attributing the variability in output quality to the cumulative predictive uncertainty along a generative path. To quantify this uncertainty, we introduce Denoising Entropy, a computable metric that serves as an internal signal for evaluating generative process. Leveraging this metric, we propose two algorithms designed to optimize the decoding path: a post-hoc selection method and a real-time guidance strategy. Experiments demonstrate that our entropy-guided methods significantly improve generation quality, consistently boosting accuracy on challenging reasoning, planning, and code benchmarks. Our work establishes Denoising Entropy as a principled tool for understanding and controlling generation, effectively turning the uncertainty in MDMs from a liability into a key advantage for discovering high-quality solutions.

Method: Propose incorporating world knowledge features including linked entities and fine-grained entity types into neural question generation models to encode additional entity-related information.

Result: World knowledge enriched model outperforms vanilla neural question generation by 1.37 BLEU-4 on SQuAD and 1.59 BLEU-4 on MS MARCO test datasets.

Conclusion: Incorporating world knowledge (entities and entity types) significantly improves question generation quality, demonstrating the value of external knowledge for generating more natural, human-like questions.

Abstract: In this paper, we propose a method for incorporating world knowledge (linked entities and fine-grained entity types) into a neural question generation model. This world knowledge helps to encode additional information related to the entities present in the passage required to generate human-like questions. We evaluate our models on both SQuAD and MS MARCO to demonstrate the usefulness of the world knowledge features. The proposed world knowledge enriched question generation model is able to outperform the vanilla neural question generation model by 1.37 and 1.59 absolute BLEU 4 score on SQuAD and MS MARCO test dataset respectively.

Method: Replaces binary MLE reward modeling with sequence MLE, enabling richer language feedback without additional annotations, models, or training stages. Uses seq2seq reward modeling to provide fine-grained language feedback.

Result: Reduces refusal-to-response paradigm in safety dialogues and long-response bias in summarization. Improves RLHF performance across 2B and 7B LLMs on 3 NLP tasks with average win rate of 76.9%. Works even under out-of-distribution prompts.

Conclusion: Seq2seq reward modeling effectively mitigates RLHF’s biased local optimization by providing richer language feedback, improving alignment without additional resources.

Abstract: Aligning the behavior of Large language models (LLMs) with human intentions and values remains a critical challenge. Reinforcement learning from human feedback (RLHF) aligns LLMs by training a reward model (RM) on human preferences and fine-tuning the LLMs to maximize RM feedback. Despite its effectiveness and popularity, RLHF is prone to biased local optimization. It means RM fails to provide feedback that accurately aligns with human preference, causing LLMs to explore unexpected generalizations, and failing to achieve alignment objectives. To mitigate this issue, we propose a novel \textit{sequence-to-sequence (seq2seq) reward modeling} method. Its key insight is that learning from language feedback rather than scalar feedback improves RLHF without additional annotations. We replaced the reward modeling target from binary maximum likelihood estimation (MLE) with sequence MLE. This method enables richer and fine-grained language feedback without additional annotations, models, or training stages. Our experiments demonstrated its effectiveness, specifically, reducing the refusal-to-response paradigm in single-turn safety dialogues and the long-response bias in text summarization tasks. We provide further analysis that seq2seq RM improves RLHF performance across 2B and 7B LLMs on 3 NLP tasks, achieving an average win rate of 76.9%. We further show that seq2seq RM can still improve the performance of RLHF under out-of-distribution prompts.

Method: CAKE frames KV cache eviction as a “cake-slicing problem,” assesses layer-specific preferences by analyzing attention dynamics in spatial and temporal dimensions, allocates cache sizes accordingly, and manages memory constraints in a cascading manner with a new eviction indicator that considers shifting token importance over time.

Result: CAKE maintains model performance with only 3.2% of KV cache, consistently outperforms current baselines across various models and memory constraints (especially in low-memory settings), and achieves over 10x speedup in decoding latency for 128K token contexts with FlashAttention-2.

Conclusion: CAKE provides an effective solution for KV cache management that adaptively allocates resources across layers while maintaining memory budgets, significantly reducing inference burden while preserving model performance.

Abstract: Large language models (LLMs) excel at processing long sequences, boosting demand for key-value (KV) caching. While recent efforts to evict KV cache have alleviated the inference burden, they often fail to allocate resources rationally across layers with different attention patterns. In this paper, we introduce Cascading and Adaptive KV cache Eviction (CAKE), a novel approach that frames KV cache eviction as a “cake-slicing problem.” CAKE assesses layer-specific preferences by considering attention dynamics in both spatial and temporal dimensions, allocates rational cache size for layers accordingly, and manages memory constraints in a cascading manner. This approach enables a global view of cache allocation, adaptively distributing resources across diverse attention mechanisms while maintaining memory budgets. CAKE also employs a new eviction indicator that considers the shifting importance of tokens over time, addressing limitations in existing methods that overlook temporal dynamics. Comprehensive experiments on LongBench and NeedleBench show that CAKE maintains model performance with only 3.2% of the KV cache and consistently outperforms current baselines across various models and memory constraints, particularly in low-memory settings. Additionally, CAKE achieves over 10x speedup in decoding latency compared to full cache when processing contexts of 128K tokens with FlashAttention-2. Our code is available at https://github.com/antgroup/cakekv.

Method: Two-part approach: (1) Fine-tune a compact 8B-parameter model with teacher-generated reasoning traces for real-time breakdown detection and explanation; (2) Systematically evaluate frontier LLMs with advanced prompting (few-shot, chain-of-thought, analogical reasoning) for high-fidelity assessment. These are integrated into an escalation architecture where the efficient detector defers to larger models only when necessary.

Result: The fine-tuned model achieves robust classification and calibration on English and Japanese dialogues, generalizes to BETOLD dataset with 7% accuracy improvement over baseline. Achieves state-of-the-art performance on DBDC5 and strong results on BETOLD, outperforming specialized classifiers. The escalation pipeline reduces inference costs by 54%.

Conclusion: The proposed “Detect, Explain, Escalate” framework provides a cost-effective and interpretable solution for robust conversational AI, balancing performance with operational efficiency through intelligent resource allocation between compact and frontier models.

Abstract: Large Language Models (LLMs) have demonstrated substantial capabilities in conversational AI applications, yet their susceptibility to dialogue breakdowns poses significant challenges to deployment reliability and user trust. This paper introduces a “Detect, Explain, Escalate” framework to manage dialogue breakdowns in LLM-powered agents, emphasizing resource-efficient operation. Our approach integrates two key strategies: (1) We fine-tune a compact 8B-parameter model, augmented with teacher-generated reasoning traces, which serves as an efficient real-time breakdown detector and explainer. This model demonstrates robust classification and calibration on English and Japanese dialogues, and generalizes to the BETOLD dataset, improving accuracy by 7% over its baseline. (2) We systematically evaluate frontier LLMs using advanced prompting (few-shot, chain-of-thought, analogical reasoning) for high-fidelity breakdown assessment. These are integrated into an “escalation” architecture where our efficient detector defers to larger models only when necessary, substantially reducing operational costs and computational overhead. Our fine-tuned model and prompting strategies achieve state-of-the-art performance on DBDC5 and strong results on BETOLD, outperforming specialized classifiers on DBDC5 and narrowing the performance gap to larger proprietary models. The proposed monitor-escalate pipeline reduces inference costs by 54%, providing a cost-effective and interpretable solution for robust conversational AI in high-impact domains. Code and models will be publicly released.

Method: The authors categorize memory into parametric (implicit in model weights) and contextual (explicit external data, structured/unstructured) forms, then define six core memory operations: Consolidation, Updating, Indexing, Forgetting, Retrieval, and Condensation.

Result: The taxonomy reveals four key research topics: long-term memory, long-context memory, parametric modification, and multi-source memory. It provides a structured framework for understanding memory-related research, benchmarks, and tools in LLM-based agents.

Conclusion: This systematic taxonomy clarifies functional interactions in LLM-based agents’ memory systems, guides future research advancements, and provides publicly available resources including datasets, papers, and tools.

Abstract: Memory is fundamental to large language model (LLM)-based agents, but existing surveys emphasize application-level use (e.g., personalized dialogue), while overlooking the atomic operations governing memory dynamics. This work categorizes memory into parametric (implicit in model weights) and contextual (explicit external data, structured/unstructured) forms, and defines six core operations: Consolidation, Updating, Indexing, Forgetting, Retrieval, and Condensation. Mapping these dimensions reveals four key research topics: long-term, long-context, parametric modification, and multi-source memory. The taxonomy provides a structured view of memory-related research, benchmarks, and tools, clarifying functional interactions in LLM-based agents and guiding future advancements. The datasets, papers, and tools are publicly available at https://github.com/Elvin-Yiming-Du/Survey_Memory_in_AI.

Method: Proposes Prune-on-Logic, a structure-aware framework that transforms Long-CoT reasoning into logic graphs and selectively prunes low-utility reasoning steps under self-verification constraints. Tests three pruning strategies: entire chains, core reasoning, and verification pruning.

Result: Verification pruning consistently improves accuracy while reducing token usage, whereas pruning reasoning steps or indiscriminate pruning degrades performance. Larger models benefit more from pruning due to richer but more redundant reasoning. Gains hold across tasks, model scales, and CoT capability.

Conclusion: Effective pruning aligns supervision with model capacity rather than merely shortening inputs. Pruning serves as a structural optimization strategy for aligning CoT reasoning with SLM capacity, with verification pruning being particularly effective.

Abstract: Long chain-of-thought (Long-CoT) reasoning improves accuracy in LLMs, yet its verbose, self-reflective style often hinders effective distillation into small language models (SLMs). We revisit Long-CoT compression through the lens of capability alignment and ask: Can pruning improve reasoning? We propose Prune-on-Logic, a structure-aware framework that transforms Long-CoT into logic graphs and selectively prunes low-utility reasoning steps under self-verification constraints. Through systematic analysis across three pruning strategies targeting entire chains, core reasoning, and verification, we find that verification pruning consistently improves accuracy while reducing token usage, whereas pruning reasoning steps or indiscriminate pruning degrades performance. Our study reveals that effective pruning aligns supervision with model capacity rather than merely shortening inputs. Gains hold across tasks, model scales, and CoT capability, with larger models benefiting more from pruning due to richer but more redundant reasoning. Our empirical findings highlight pruning as a structural optimization strategy for aligning CoT reasoning with SLM capacity.

Method: Systematically evaluated Qwen3 (1.7B-235B) and DeepSeek-R1 (1.5B-70B) models across 15 medical datasets with controlled thinking budgets ranging from zero to unlimited tokens.

Result: Found logarithmic scaling relationships between accuracy and thinking budget/model size, identified three efficiency regimes (0-256, 256-512, 512+ tokens), and discovered smaller models benefit disproportionately more from extended thinking (15-20% vs 5-10% improvements).

Conclusion: Thinking budget control is critical for optimizing medical AI systems, enabling dynamic resource allocation aligned with clinical needs while maintaining transparency for healthcare deployment.

Abstract: This study presents the first comprehensive evaluation of thinking budget mechanisms in medical reasoning tasks, revealing fundamental scaling laws between computational resources and reasoning quality. We systematically evaluated two major model families, Qwen3 (1.7B to 235B parameters) and DeepSeek-R1 (1.5B to 70B parameters), across 15 medical datasets spanning diverse specialties and difficulty levels. Through controlled experiments with thinking budgets ranging from zero to unlimited tokens, we establish logarithmic scaling relationships where accuracy improvements follow a predictable pattern with both thinking budget and model size. Our findings identify three distinct efficiency regimes: high-efficiency (0 to 256 tokens) suitable for real-time applications, balanced (256 to 512 tokens) offering optimal cost-performance tradeoffs for routine clinical support, and high-accuracy (above 512 tokens) justified only for critical diagnostic tasks. Notably, smaller models demonstrate disproportionately larger benefits from extended thinking, with 15 to 20% improvements compared to 5 to 10% for larger models, suggesting a complementary relationship where thinking budget provides greater relative benefits for capacity-constrained models. Domain-specific patterns emerge clearly, with neurology and gastroenterology requiring significantly deeper reasoning processes than cardiovascular or respiratory medicine. The consistency between Qwen3 native thinking budget API and our proposed truncation method for DeepSeek-R1 validates the generalizability of thinking budget concepts across architectures. These results establish thinking budget control as a critical mechanism for optimizing medical AI systems, enabling dynamic resource allocation aligned with clinical needs while maintaining the transparency essential for healthcare deployment.

Method: GAICo provides a unified, extensible framework with comprehensive reference-based metrics for unstructured text, specialized structured data formats, and multimedia (images, audio). It features a high-level API for end-to-end analysis (multi-model comparison, visualization, reporting) and direct metric access for granular control.

Result: GAICo demonstrates utility through a detailed case study evaluating complex multi-modal AI Travel Assistant pipelines. The tool has been downloaded over 13K times across versions since its PyPI release in June 2025, showing growing community adoption.

Conclusion: GAICo empowers AI researchers and developers to efficiently assess system performance, make evaluation reproducible, improve development velocity, and build more trustworthy AI systems, aligning with the goal of moving faster and safer in AI deployment.

Abstract: The rapid proliferation of Generative AI (GenAI) into diverse, high-stakes domains necessitates robust and reproducible evaluation methods. However, practitioners often resort to ad-hoc, non-standardized scripts, as common metrics are often unsuitable for specialized, structured outputs (e.g., automated plans, time-series) or holistic comparison across modalities (e.g., text, audio, and image). This fragmentation hinders comparability and slows AI system development. To address this challenge, we present GAICo (Generative AI Comparator): a deployed, open-source Python library that streamlines and standardizes GenAI output comparison. GAICo provides a unified, extensible framework supporting a comprehensive suite of reference-based metrics for unstructured text, specialized structured data formats, and multimedia (images, audio). Its architecture features a high-level API for rapid, end-to-end analysis, from multi-model comparison to visualization and reporting, alongside direct metric access for granular control. We demonstrate GAICo’s utility through a detailed case study evaluating and debugging complex, multi-modal AI Travel Assistant pipelines. GAICo empowers AI researchers and developers to efficiently assess system performance, make evaluation reproducible, improve development velocity, and ultimately build more trustworthy AI systems, aligning with the goal of moving faster and safer in AI deployment. Since its release on PyPI in Jun 2025, the tool has been downloaded over 13K times, across versions, by Aug 2025, demonstrating growing community interest.

Method: Proposes unsupervised adaptation of LLMs using context compression as pretext task. Models learn to generate compact memory tokens that substitute entire context for downstream prediction. Combines compression with contrastive learning for further improvements.

Result: LLM2Comp outperforms contemporary LLM-based text encoders across wide range of tasks. Shows better sample efficiency, requiring significantly less training data than token-level approaches like LLM2Vec.

Conclusion: Context compression is an effective pretext task for adapting LLMs to text representation, superior to token-level objectives. The approach produces strong, sample-efficient representation models.

Abstract: Text representation plays a critical role in tasks like clustering, retrieval, and other downstream applications. With the emergence of large language models (LLMs), there is increasing interest in harnessing their capabilities for this purpose. However, most of the LLMs are inherently causal and optimized for next-token prediction, making them suboptimal for producing holistic representations. To address this, recent studies introduced pretext tasks to adapt LLMs for text representation. Most of these tasks, however, rely on token-level prediction objectives, such as the masked next-token prediction (MNTP) used in LLM2Vec. In this work, we explore the untapped potential of context compression as a pretext task for unsupervised adaptation of LLMs. During compression pre-training, the model learns to generate compact memory tokens, which substitute the whole context for downstream sequence prediction. Experiments demonstrate that a well-designed compression objective can significantly enhance LLM-based text representations, outperforming models trained with token-level pretext tasks. Further improvements through contrastive learning produce a strong representation model (LLM2Comp) that outperforms contemporary LLM-based text encoders on a wide range of tasks while being more sample-efficient, requiring significantly less training data. Code is available at https://github.com/longtaizi13579/LLM2Comp.

Method: Created a benchmark with 2,800 carefully curated Chinese medical exam questions across 7 specialties (cardiovascular, gastroenterology, hematology, infectious diseases, nephrology, neurology, respiratory) at two difficulty levels (attending vs senior physician). Evaluated 27 state-of-the-art LLMs using this framework.

Result: Mixtral-8x7B achieved highest overall accuracy (74.25%), followed by DeepSeek-R1-671B (64.07%). No consistent correlation between model size and performance. Performance varied significantly by specialty (better on cardiovascular/neurology vs gastroenterology/nephrology). Top models showed minimal performance degradation between difficulty levels.

Conclusion: LLMs show promise for medical applications but have limitations in specialized contexts. Mixture-of-experts architectures can outperform larger models. The benchmark provides critical insights for deploying LLMs in medical education and clinical decision support systems.

Abstract: The rapid advancement of large language models(LLMs) has prompted significant interest in their potential applications in medical domains. This paper presents a comprehensive benchmark evaluation of 27 state-of-the-art LLMs on Chinese medical examination questions, encompassing seven medical specialties across two professional levels. We introduce a robust evaluation framework that assesses model performance on 2,800 carefully curated questions from cardiovascular, gastroenterology, hematology, infectious diseases, nephrology, neurology, and respiratory medicine domains. Our dataset distinguishes between attending physician and senior physician difficulty levels, providing nuanced insights into model capabilities across varying complexity. Our empirical analysis reveals substantial performance variations among models, with Mixtral-8x7B achieving the highest overall accuracy of 74.25%, followed by DeepSeek-R1-671B at 64.07%. Notably, we observe no consistent correlation between model size and performance, as evidenced by the strong performance of smaller mixture-of-experts architectures. The evaluation demonstrates significant performance gaps between medical specialties, with models generally performing better on cardiovascular and neurology questions compared to gastroenterology and nephrology domains. Furthermore, our analysis indicates minimal performance degradation between attending and senior physician levels for top-performing models, suggesting robust generalization capabilities. This benchmark provides critical insights for the deployment of LLMs in medical education and clinical decision support systems, highlighting both the promise and current limitations of these technologies in specialized medical contexts.

Method: Tournament-style ELO ranking with multi-agent reasoning, configurable tournament parameters, dynamic agent selection, and multiple consensus fusion strategies.

Result: 8.4% improvement in overall quality metrics, R^2 values exceeding 0.96 in ELO rating convergence, and significant improvements in accuracy, coherence, and reliability.

Conclusion: ART provides a scalable, production-ready solution for applications requiring high-quality, vetted LLM responses through systematic multi-agent collaboration.

Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language understanding and generation. However, single-model responses often exhibit inconsistencies, hallucinations, and varying quality across different query domains. This paper presents ART (Adaptive Response Tuning), a novel framework that employs tournament-style ELO ranking and multi-agent reasoning to systematically optimize LLM outputs. By enabling multiple LLM agents to compete, critique, and collaborate through structured tournament workflows, ART produces consensus responses that outperform individual model outputs. Our framework introduces configurable tournament parameters, dynamic agent selection, and multiple consensus fusion strategies. Experimental evaluations demonstrate significant improvements in response accuracy, coherence, and reliability compared to baseline single-model approaches. The ART framework provides a scalable, production-ready solution for applications requiring high-quality, vetted LLM responses, achieving an 8.4% improvement in overall quality metrics and R^2 values exceeding 0.96 in ELO rating convergence.

Method: Created VLegal-Bench using Bloom’s cognitive taxonomy to design tasks reflecting practical legal usage scenarios. Developed a rigorous annotation pipeline where legal experts label and cross-validate 10,450 samples, ensuring each is grounded in authoritative legal documents. The benchmark includes general legal Q&A, retrieval-augmented generation, multi-step reasoning, and scenario-based problem solving tailored to Vietnamese law.

Result: Established the first comprehensive benchmark for Vietnamese legal LLM evaluation with 10,450 expert-annotated samples. Created a standardized, transparent, and cognitively informed evaluation framework that supports assessment of LLM performance in Vietnamese legal contexts.

Conclusion: VLegal-Bench provides a solid foundation for assessing LLM performance in Vietnamese legal contexts and supports the development of more reliable, interpretable, and ethically aligned AI-assisted legal systems. The benchmark is publicly available at https://vilegalbench.cmcai.vn/ to facilitate access and reproducibility.

Abstract: The rapid advancement of large language models (LLMs) has enabled new possibilities for applying artificial intelligence within the legal domain. Nonetheless, the complexity, hierarchical organization, and frequent revisions of Vietnamese legislation pose considerable challenges for evaluating how well these models interpret and utilize legal knowledge. To address this gap, the Vietnamese Legal Benchmark (VLegal-Bench) is introduced, the first comprehensive benchmark designed to systematically assess LLMs on Vietnamese legal tasks. Informed by Bloom’s cognitive taxonomy, VLegal-Bench encompasses multiple levels of legal understanding through tasks designed to reflect practical usage scenarios. The benchmark comprises 10,450 samples generated through a rigorous annotation pipeline, where legal experts label and cross-validate each instance using our annotation system to ensure every sample is grounded in authoritative legal documents and mirrors real-world legal assistant workflows, including general legal questions and answers, retrieval-augmented generation, multi-step reasoning, and scenario-based problem solving tailored to Vietnamese law. By providing a standardized, transparent, and cognitively informed evaluation framework, VLegal-Bench establishes a solid foundation for assessing LLM performance in Vietnamese legal contexts and supports the development of more reliable, interpretable, and ethically aligned AI-assisted legal systems. To facilitate access and reproducibility, we provide a public landing page for this benchmark at https://vilegalbench.cmcai.vn/.

Method: Adapts pretrained Gemma 3 decoder-only models into encoder-decoder architecture using UL2 adaptation recipe, with two efficiency improvements: tied word embeddings (shared across encoder/decoder) and merged attention (unifying decoder self- and cross-attention).

Result: Demonstrates generality of adaptation strategy across architectures/modalities, shows encoder-decoder strength in long-context modeling, achieves comparable/better pretraining performance and significantly improved post-training performance compared to Gemma 3 counterparts.

Conclusion: T5Gemma 2 successfully extends the adaptation approach to multimodal models with efficiency improvements, releasing pretrained models (270M-270M, 1B-1B, 4B-4B) for community research.

Abstract: We introduce T5Gemma 2, the next generation of the T5Gemma family of lightweight open encoder-decoder models, featuring strong multilingual, multimodal and long-context capabilities. T5Gemma 2 follows the adaptation recipe (via UL2) in T5Gemma – adapting a pretrained decoder-only model into an encoder-decoder model, and extends it from text-only regime to multimodal based on the Gemma 3 models. We further propose two methods to improve the efficiency: tied word embedding that shares all embeddings across encoder and decoder, and merged attention that unifies decoder self- and cross-attention into a single joint module. Experiments demonstrate the generality of the adaptation strategy over architectures and modalities as well as the unique strength of the encoder-decoder architecture on long context modeling. Similar to T5Gemma, T5Gemma 2 yields comparable or better pretraining performance and significantly improved post-training performance than its Gemma 3 counterpart. We release the pretrained models (270M-270M, 1B-1B and 4B-4B) to the community for future research.

Method: The paper introduces an evaluation framework that reports boundary density and segment alignment diagnostics (purity and coverage) alongside window-tolerant F1 (W-F1). This separates boundary scoring from boundary selection, allowing evaluation across different density regimes rather than at a single operating point. The framework is tested across eight dialogue datasets spanning task-oriented, open-domain, meeting-style, and synthetic interactions.

Result: Cross-dataset evaluation shows that reported performance differences often reflect annotation granularity mismatch rather than boundary placement quality alone. Boundary-based metrics are strongly coupled to boundary density, with threshold sweeps producing larger W-F1 changes than switching between segmentation methods.

Conclusion: Topic segmentation should be viewed as a granularity selection problem rather than prediction of a single correct boundary set. This motivates separating boundary scoring from boundary selection for analyzing and tuning segmentation under varying annotation granularities, providing a more nuanced evaluation framework for modern conversational systems.

Abstract: Dialogue topic segmentation supports summarization, retrieval, memory management, and conversational continuity. Despite decades of work, evaluation practice remains dominated by strict boundary matching and F1-based metrics. Modern large language model (LLM) based conversational systems increasingly rely on segmentation to manage conversation history beyond fixed context windows. In such systems, unstructured context accumulation degrades efficiency and coherence. This paper introduces an evaluation framework that reports boundary density and segment alignment diagnostics (purity and coverage) alongside window-tolerant F1 (W-F1). By separating boundary scoring from boundary selection, we evaluate segmentation quality across density regimes rather than at a single operating point. Cross-dataset evaluation shows that reported performance differences often reflect annotation granularity mismatch rather than boundary placement quality alone. We evaluate structurally distinct segmentation strategies across eight dialogue datasets spanning task-oriented, open-domain, meeting-style, and synthetic interactions. Boundary-based metrics are strongly coupled to boundary density: threshold sweeps produce larger W-F1 changes than switching between methods. These findings support viewing topic segmentation as a granularity selection problem rather than prediction of a single correct boundary set. This motivates separating boundary scoring from boundary selection for analyzing and tuning segmentation under varying annotation granularities.

Method: Proposes a human-centered framework with three dimensions: augmented terminologist (AI as capability amplifier), ethical AI, and human-centered design, emphasizing human control and terminologist centrality.

Result: A framework that conceptualizes AI as amplifying terminologists’ capabilities rather than replacing them, ensuring compatibility of high automation with strong human control and bias mitigation.

Conclusion: Current AI adoption choices will shape terminological practice and the preservation of accuracy, adequacy, and diversity in terminology and specialized knowledge.

Abstract: The rapid diffusion of generative artificial intelligence is transforming terminology work. While this technology promises gains in efficiency, its unstructured adoption risks weakening professional autonomy, amplifying bias, and eroding linguistic and conceptual diversity. This paper argues that a human-centered approach to artificial intelligence has become a necessity for terminology work. Building on research in artificial intelligence and translation studies, it proposes a human-centered framework that conceptualizes artificial intelligence as a means of amplifying the terminologist’s capabilities, rather than replacing them. The framework is organized around three interrelated dimensions: the augmented terminologist, ethical AI, and human-centered design. Together, these dimensions emphasize the compatibility of high automation with strong human control, the central role of terminologists in bias mitigation, and the importance of designing AI tools and workflows around the needs, values, and well-being of the terminologist. The paper concludes by stressing that current choices in AI adoption will shape not only terminological practice, but also the preservation of accuracy, adequacy, and diversity in terminology and specialized knowledge.

Method: Proposes M³KG-RAG with two main components: 1) A lightweight multi-agent pipeline to construct multi-hop MMKGs (M³KG) with context-enriched triplets of multimodal entities, enabling modality-wise retrieval. 2) GRASP (Grounded Retrieval And Selective Pruning) that ensures precise entity grounding, evaluates answer-supporting relevance, and prunes redundant context.

Result: Extensive experiments across diverse multimodal benchmarks demonstrate that M³KG-RAG significantly enhances MLLMs’ multimodal reasoning and grounding capabilities compared to existing approaches.

Conclusion: M³KG-RAG effectively addresses limitations in multimodal RAG by improving knowledge graph construction and retrieval mechanisms, leading to better reasoning depth and answer faithfulness in audio-visual domains.

Abstract: Retrieval-Augmented Generation (RAG) has recently been extended to multimodal settings, connecting multimodal large language models (MLLMs) with vast corpora of external knowledge such as multimodal knowledge graphs (MMKGs). Despite their recent success, multimodal RAG in the audio-visual domain remains challenging due to 1) limited modality coverage and multi-hop connectivity of existing MMKGs, and 2) retrieval based solely on similarity in a shared multimodal embedding space, which fails to filter out off-topic or redundant knowledge. To address these limitations, we propose M$^3$KG-RAG, a Multi-hop Multimodal Knowledge Graph-enhanced RAG that retrieves query-aligned audio-visual knowledge from MMKGs, improving reasoning depth and answer faithfulness in MLLMs. Specifically, we devise a lightweight multi-agent pipeline to construct multi-hop MMKG (M$^3$KG), which contains context-enriched triplets of multimodal entities, enabling modality-wise retrieval based on input queries. Furthermore, we introduce GRASP (Grounded Retrieval And Selective Pruning), which ensures precise entity grounding to the query, evaluates answer-supporting relevance, and prunes redundant context to retain only knowledge essential for response generation. Extensive experiments across diverse multimodal benchmarks demonstrate that M$^3$KG-RAG significantly enhances MLLMs’ multimodal reasoning and grounding over existing approaches.

Method: Introduces Step-DeepResearch agent with: 1) Data Synthesis Strategy Based on Atomic Capabilities to reinforce planning and report writing, 2) Progressive training path from agentic mid-training to SFT and RL, 3) Checklist-style Judger for robustness, and 4) ADR-Bench for Chinese domain evaluation.

Result: Step-DeepResearch (32B) scores 61.4% on Scale AI Research Rubrics. On ADR-Bench, it significantly outperforms comparable models and rivals SOTA closed-source models like OpenAI and Gemini DeepResearch.

Conclusion: Refined training enables medium-sized models to achieve expert-level capabilities at industry-leading cost-efficiency, proving that well-designed training approaches can make smaller models competitive with larger closed-source alternatives.

Abstract: As LLMs shift toward autonomous agents, Deep Research has emerged as a pivotal metric. However, existing academic benchmarks like BrowseComp often fail to meet real-world demands for open-ended research, which requires robust skills in intent recognition, long-horizon decision-making, and cross-source verification. To address this, we introduce Step-DeepResearch, a cost-effective, end-to-end agent. We propose a Data Synthesis Strategy Based on Atomic Capabilities to reinforce planning and report writing, combined with a progressive training path from agentic mid-training to SFT and RL. Enhanced by a Checklist-style Judger, this approach significantly improves robustness. Furthermore, to bridge the evaluation gap in the Chinese domain, we establish ADR-Bench for realistic deep research scenarios. Experimental results show that Step-DeepResearch (32B) scores 61.4% on Scale AI Research Rubrics. On ADR-Bench, it significantly outperforms comparable models and rivals SOTA closed-source models like OpenAI and Gemini DeepResearch. These findings prove that refined training enables medium-sized models to achieve expert-level capabilities at industry-leading cost-efficiency.

Method: NeRV360 integrates viewport extraction into decoding and introduces a spatial-temporal affine transform module for conditional decoding based on viewpoint and time, enabling selective decoding of user-selected viewports.

Result: On 6K-resolution videos, NeRV360 achieves 7-fold reduction in memory consumption and 2.5-fold increase in decoding speed compared to HNeRV, while delivering better image quality in objective metrics.

Conclusion: NeRV360 enables practical real-time applications for 360° video compression by efficiently decoding only the necessary viewport content, significantly improving performance over existing NeRV approaches.

Abstract: Implicit neural representations for videos (NeRV) have shown strong potential for video compression. However, applying NeRV to high-resolution 360-degree videos causes high memory usage and slow decoding, making real-time applications impractical. We propose NeRV360, an end-to-end framework that decodes only the user-selected viewport instead of reconstructing the entire panoramic frame. Unlike conventional pipelines, NeRV360 integrates viewport extraction into decoding and introduces a spatial-temporal affine transform module for conditional decoding based on viewpoint and time. Experiments on 6K-resolution videos show that NeRV360 achieves a 7-fold reduction in memory consumption and a 2.5-fold increase in decoding speed compared to HNeRV, a representative prior work, while delivering better image quality in terms of objective metrics.

Method: Created VL4Gaze benchmark with 489K automatically generated QA pairs across 124K images, formulating gaze understanding as a unified VQA problem through four tasks: gaze object description, gaze direction description, gaze point location, and ambiguous question recognition.

Result: Large-scale VLMs struggle to reliably infer gaze semantics and spatial localization without task-specific supervision. Training on VL4Gaze brings substantial and consistent improvements across all tasks, highlighting the importance of targeted multi-task supervision.

Conclusion: Gaze understanding does not emerge naturally from general-purpose vision-language pre-training; targeted multi-task supervision (like VL4Gaze) is essential for developing reliable gaze interpretation capabilities in VLMs.

Abstract: Human gaze provides essential cues for interpreting attention, intention, and social interaction in visual scenes, yet gaze understanding remains largely unexplored in current vision-language models (VLMs). While recent VLMs achieve strong scene-level reasoning across a range of visual tasks, there exists no benchmark that systematically evaluates or trains them for gaze interpretation, leaving open the question of whether gaze understanding can emerge from general-purpose vision-language pre-training. To address this gap, we introduce VL4Gaze, the first large-scale benchmark designed to investigate, evaluate, and unlock the potential of VLMs for gaze understanding. VL4Gaze contains 489K automatically generated question-answer pairs across 124K images and formulates gaze understanding as a unified VQA problem through four complementary tasks: (1) gaze object description, (2) gaze direction description, (3) gaze point location, and (4) ambiguous question recognition. We comprehensively evaluate both commercial and open-source VLMs under in-context learning and fine-tuning settings. The results show that even large-scale VLMs struggle to reliably infer gaze semantics and spatial localization without task-specific supervision. In contrast, training on VL4Gaze brings substantial and consistent improvements across all tasks, highlighting the importance of targeted multi-task supervision for developing gaze understanding capabilities in VLMs. We will release the dataset and code to support further research and development in this direction.

Method: Iterative hardware-aware neural architecture search using Once-for-All-style ResDets supernet with evolutionary search alternating between backbone and neck/head optimization, plus population passthrough mechanism and accuracy predictor to reduce search cost.

Result: TrashDet family (1.2M-30.5M parameters) achieves 11.4-19.5 mAP50 on TACO subset; best variant improves accuracy by 3.6 mAP50 with fewer parameters. On MAX78002 microcontroller, specialized variants reduce energy by 88%, latency by 78%, and power by 53% while improving mAP50 by 10.2%.

Conclusion: The proposed hardware-aware NAS framework successfully generates scalable, deployment-ready trash detectors optimized for diverse TinyML hardware constraints, significantly improving efficiency-performance trade-offs for edge trash detection applications.

Abstract: This paper addresses trash detection on the TACO dataset under strict TinyML constraints using an iterative hardware-aware neural architecture search framework targeting edge and IoT devices. The proposed method constructs a Once-for-All-style ResDets supernet and performs iterative evolutionary search that alternates between backbone and neck/head optimization, supported by a population passthrough mechanism and an accuracy predictor to reduce search cost and improve stability. This framework yields a family of deployment-ready detectors, termed TrashDets. On a five-class TACO subset (paper, plastic, bottle, can, cigarette), the strongest variant, TrashDet-l, achieves 19.5 mAP50 with 30.5M parameters, improving accuracy by up to 3.6 mAP50 over prior detectors while using substantially fewer parameters. The TrashDet family spans 1.2M to 30.5M parameters with mAP50 values between 11.4 and 19.5, providing scalable detector options for diverse TinyML deployment budgets on resource-constrained hardware. On the MAX78002 microcontroller with the TrashNet dataset, two specialized variants, TrashDet-ResNet and TrashDet-MBNet, jointly dominate the ai87-fpndetector baseline, with TrashDet-ResNet achieving 7525~$μ$J energy per inference at 26.7 ms latency and 37.45 FPS, and TrashDet-MBNet improving mAP50 by 10.2%; together they reduce energy consumption by up to 88%, latency by up to 78%, and average power by up to 53% compared to existing TinyML detectors.

Method: Introduces OccuFly benchmark captured at 50m, 40m, and 30m altitudes across four seasons. Proposes LiDAR-free data generation framework using camera modality with traditional 3D reconstruction. Automates label transfer by lifting annotated 2D masks into reconstructed point clouds to minimize manual 3D annotation effort.

Result: Created first real-world camera-based aerial SSC benchmark covering urban, industrial, and rural scenarios with 22 semantic classes. Data format adheres to established conventions for integration with existing research. Benchmarked state-of-the-art methods and identified challenges specific to elevated viewpoints.

Conclusion: OccuFly provides a comprehensive vision benchmark for holistic aerial 3D scene understanding, addressing the gap in aerial SSC research and enabling progress on downstream applications for autonomous flying systems.

Abstract: Semantic Scene Completion (SSC) is crucial for 3D perception in mobile robotics, as it enables holistic scene understanding by jointly estimating dense volumetric occupancy and per-voxel semantics. Although SSC has been widely studied in terrestrial domains such as autonomous driving, aerial scenarios like autonomous flying remain largely unexplored, thereby limiting progress on downstream applications. Furthermore, LiDAR sensors represent the primary modality for SSC data generation, which poses challenges for most uncrewed aerial vehicles (UAVs) due to flight regulations, mass and energy constraints, and the sparsity of LiDAR-based point clouds from elevated viewpoints. To address these limitations, we introduce OccuFly, the first real-world, camera-based aerial SSC benchmark, captured at altitudes of 50m, 40m, and 30m during spring, summer, fall, and winter. OccuFly covers urban, industrial, and rural scenarios, provides 22 semantic classes, and the data format adheres to established conventions to facilitate seamless integration with existing research. Crucially, we propose a LiDAR-free data generation framework based on camera modality, which is ubiquitous on modern UAVs. By utilizing traditional 3D reconstruction, our framework automates label transfer by lifting a subset of annotated 2D masks into the reconstructed point cloud, thereby substantially minimizing manual 3D annotation effort. Finally, we benchmark the state-of-the-art on OccuFly and highlight challenges specific to elevated viewpoints, yielding a comprehensive vision benchmark for holistic aerial 3D scene understanding.

Method: Proposes NullBUS framework with nullable prompts implemented as learnable null embeddings with presence masks, enabling fallback to image-only evidence when metadata is absent while still utilizing text when available, all in a single model.

Result: Achieves mean IoU of 0.8568 and mean Dice of 0.9103 on a unified pool of three public BUS datasets, demonstrating state-of-the-art performance under mixed prompt availability conditions.

Conclusion: NullBUS effectively addresses the challenge of missing metadata in public BUS datasets by enabling mixed-supervision learning, improving segmentation robustness and performance across datasets with varying prompt availability.

Abstract: Breast ultrasound (BUS) segmentation provides lesion boundaries essential for computer-aided diagnosis and treatment planning. While promptable methods can improve segmentation performance and tumor delineation when text or spatial prompts are available, many public BUS datasets lack reliable metadata or reports, constraining training to small multimodal subsets and reducing robustness. We propose NullBUS, a multimodal mixed-supervision framework that learns from images with and without prompts in a single model. To handle missing text, we introduce nullable prompts, implemented as learnable null embeddings with presence masks, enabling fallback to image-only evidence when metadata are absent and the use of text when present. Evaluated on a unified pool of three public BUS datasets, NullBUS achieves a mean IoU of 0.8568 and a mean Dice of 0.9103, demonstrating state-of-the-art performance under mixed prompt availability.

Method: Task-driven co-design framework unifying optics, sensor modeling, and lightweight semantic segmentation networks into single end-to-end RAW-to-task pipeline. Integrates realistic cellphone-scale lens models, learnable color filter arrays, Poisson-Gaussian noise processes, and quantization, all optimized directly for segmentation objectives.

Result: Consistent mIoU improvements on KITTI-360 over fixed pipelines, with optics modeling and CFA learning providing largest gains (especially for thin or low-light-sensitive classes). Achieved with compact ~1M-parameter model running at ~28 FPS, demonstrating edge deployability.

Conclusion: Full-stack co-optimization of optics, sensors, and networks establishes principled path toward efficient, reliable, and deployable perception in autonomous systems, with co-designed sensors adapting acquisition to semantic structure.

Abstract: Traditional autonomous driving pipelines decouple camera design from downstream perception, relying on fixed optics and handcrafted ISPs that prioritize human viewable imagery rather than machine semantics. This separation discards information during demosaicing, denoising, or quantization, while forcing models to adapt to sensor artifacts. We present a task-driven co-design framework that unifies optics, sensor modeling, and lightweight semantic segmentation networks into a single end-to-end RAW-to-task pipeline. Building on DeepLens[19], our system integrates realistic cellphone-scale lens models, learnable color filter arrays, Poisson-Gaussian noise processes, and quantization, all optimized directly for segmentation objectives. Evaluations on KITTI-360 show consistent mIoU improvements over fixed pipelines, with optics modeling and CFA learning providing the largest gains, especially for thin or low-light-sensitive classes. Importantly, these robustness gains are achieved with a compact ~1M-parameter model running at ~28 FPS, demonstrating edge deployability. Visual and quantitative analyses further highlight how co-designed sensors adapt acquisition to semantic structure, sharpening boundaries and maintaining accuracy under blur, noise, and low bit-depth. Together, these findings establish full-stack co-optimization of optics, sensors, and networks as a principled path toward efficient, reliable, and deployable perception in autonomous systems.

Method: Created CHAMMI-75 dataset by curating heterogeneous, multi-channel microscopy images from 75 diverse biological studies from publicly available sources to investigate channel-adaptive models.

Result: Training with CHAMMI-75 improves performance in multi-channel bioimaging tasks due to its high diversity in microscopy modalities, enabling models to process any microscopy image type.

Conclusion: CHAMMI-75 paves the way for next-generation cellular morphology models that are channel-adaptive and can be reused across diverse biological studies, overcoming current limitations of specialized models.

Abstract: Quantifying cell morphology using images and machine learning has proven to be a powerful tool to study the response of cells to treatments. However, models used to quantify cellular morphology are typically trained with a single microscopy imaging type. This results in specialized models that cannot be reused across biological studies because the technical specifications do not match (e.g., different number of channels), or because the target experimental conditions are out of distribution. Here, we present CHAMMI-75, an open access dataset of heterogeneous, multi-channel microscopy images from 75 diverse biological studies. We curated this resource from publicly available sources to investigate cellular morphology models that are channel-adaptive and can process any microscopy image type. Our experiments show that training with CHAMMI-75 can improve performance in multi-channel bioimaging tasks primarily because of its high diversity in microscopy modalities. This work paves the way to create the next generation of cellular morphology models for biological studies.

Method: TVP first solves problems using basic tools while accumulating solutions in an Example Library, then abstracts recurring patterns into reusable higher-level tools for an evolving Tool Library.

Result: Achieves SOTA on Omni3D-Bench (outperforms GPT-4o by 22% and previous best visual programming by 11%), with learned tools used 5x more frequently than inductive ones and strong generalization to unseen spatial tasks.

Conclusion: Experience-driven transductive tool creation is a powerful paradigm for building self-evolving visual programming agents that effectively tackle challenging spatial reasoning tasks.

Abstract: Spatial reasoning in 3D scenes requires precise geometric calculations that challenge vision-language models. Visual programming addresses this by decomposing problems into steps calling specialized tools, yet existing methods rely on either fixed toolsets or speculative tool induction before solving problems, resulting in suboptimal programs and poor utilization of induced tools. We present Transductive Visual Programming (TVP), a novel framework that builds new tools from its own experience rather than speculation. TVP first solves problems using basic tools while accumulating experiential solutions into an Example Library, then abstracts recurring patterns from these programs into reusable higher-level tools for an evolving Tool Library. This allows TVP to tackle new problems with increasingly powerful tools learned from experience. On Omni3D-Bench, TVP achieves state-of-the-art performance, outperforming GPT-4o by 22% and the previous best visual programming system by 11%. Our transductively learned tools are used 5x more frequently as core program dependency than inductively created ones, demonstrating more effective tool discovery and reuse. The evolved tools also show strong generalization to unseen spatial tasks, achieving superior performance on benchmarks from SpatialScore-Hard collection without any testset-specific modification. Our work establishes experience-driven transductive tool creation as a powerful paradigm for building self-evolving visual programming agents that effectively tackle challenging spatial reasoning tasks. We release our code at https://transductive-visualprogram.github.io/.

Method: Leverages TinyML techniques, wireless communication framework, and microcontroller platforms to develop a multimodal sensing system that fuses accelerometer data and vision inputs for image classification, object detection, and behavior recognition tasks.

Result: Achieves 270× model size reduction, less than 80ms response latency, and on-par performance compared to existing methods, with successful deployment on commercial microcontrollers for real-time inference.

Conclusion: Delivers a robust, scalable IoT-edge livestock monitoring solution adaptable to diverse farming needs, offering flexibility for future extensions through wireless communication capabilities.

Abstract: The advancement of technology has revolutionized the agricultural industry, transitioning it from labor-intensive farming practices to automated, AI-powered management systems. In recent years, more intelligent livestock monitoring solutions have been proposed to enhance farming efficiency and productivity. This work presents a novel approach to animal activity recognition and movement tracking, leveraging tiny machine learning (TinyML) techniques, wireless communication framework, and microcontroller platforms to develop an efficient, cost-effective livestock sensing system. It collects and fuses accelerometer data and vision inputs to build a multimodal network for three tasks: image classification, object detection, and behavior recognition. The system is deployed and evaluated on commercial microcontrollers for real-time inference using embedded applications, demonstrating up to 270$\times$ model size reduction, less than 80ms response latency, and on-par performance comparable to existing methods. The incorporation of the wireless communication technique allows for seamless data transmission between devices, benefiting use cases in remote locations with poor Internet connectivity. This work delivers a robust, scalable IoT-edge livestock monitoring solution adaptable to diverse farming needs, offering flexibility for future extensions.

Method: Proposes adaptive visual preprocessing combining content-aware image analysis, adaptive resolution selection, and content-aware cropping to reduce visual redundancy before vision encoding. Integrated with FastVLM without modifying architecture or retraining.

Result: Reduces per-image inference time by over 50%, lowers mean full generation time, and achieves consistent reduction of more than 55% in visual token count compared to baseline pipeline on DocVQA subset.

Conclusion: Input-aware preprocessing is an effective and lightweight strategy for improving deployment-oriented efficiency of vision-language models, demonstrating significant computational savings without architectural changes.

Abstract: Vision-Language Models (VLMs) have demonstrated strong performance on multimodal reasoning tasks, but their deployment remains challenging due to high inference latency and computational cost, particularly when processing high-resolution visual inputs. While recent architectures such as FastVLM improve efficiency through optimized vision encoders, existing pipelines still rely on static visual preprocessing, leading to redundant computation for visually simple inputs. In this work, we propose an adaptive visual preprocessing method that dynamically adjusts input resolution and spatial coverage based on image content characteristics. The proposed approach combines content-aware image analysis, adaptive resolution selection, and content-aware cropping to reduce visual redundancy prior to vision encoding. Importantly, the method is integrated with FastVLM without modifying its architecture or requiring retraining. We evaluate the proposed method on a subset of the DocVQA dataset in an inference-only setting, focusing on efficiency-oriented metrics. Experimental results show that adaptive preprocessing reduces per-image inference time by over 50%, lowers mean full generation time, and achieves a consistent reduction of more than 55% in visual token count compared to the baseline pipeline. These findings demonstrate that input-aware preprocessing is an effective and lightweight strategy for improving deployment-oriented efficiency of vision-language models. To facilitate reproducibility, our implementation is provided as a fork of the FastVLM repository, incorporating the files for the proposed method, and is available at https://github.com/kmdavidds/mlfastlm.

Method: ALIVE operates fully on local hardware with three key components: (1) Avatar-delivered lectures using ASR transcription, LLM refinement, and neural talking-head synthesis; (2) Content-aware retrieval combining semantic similarity with timestamp alignment; (3) Real-time multimodal interaction allowing students to pause, ask questions via text/voice, and receive avatar/text responses.

Result: Demonstrated on a complete medical imaging course with evaluation showing accurate retrieval, good latency characteristics, and positive user experience. The system provides accurate, content-aware, and engaging real-time support.

Conclusion: ALIVE shows how multimodal AI combined with content-aware retrieval and local deployment can significantly enhance recorded lectures’ pedagogical value, offering an extensible pathway toward next-generation interactive learning environments.

Abstract: Traditional lecture videos offer flexibility but lack mechanisms for real-time clarification, forcing learners to search externally when confusion arises. Recent advances in large language models and neural avatars provide new opportunities for interactive learning, yet existing systems typically lack lecture awareness, rely on cloud-based services, or fail to integrate retrieval and avatar-delivered explanations in a unified, privacy-preserving pipeline. We present ALIVE, an Avatar-Lecture Interactive Video Engine that transforms passive lecture viewing into a dynamic, real-time learning experience. ALIVE operates fully on local hardware and integrates (1) Avatar-delivered lecture generated through ASR transcription, LLM refinement, and neural talking-head synthesis; (2) A content-aware retrieval mechanism that combines semantic similarity with timestamp alignment to surface contextually relevant lecture segments; and (3) Real-time multimodal interaction, enabling students to pause the lecture, ask questions through text or voice, and receive grounded explanations either as text or as avatar-delivered responses. To maintain responsiveness, ALIVE employs lightweight embedding models, FAISS-based retrieval, and segmented avatar synthesis with progressive preloading. We demonstrate the system on a complete medical imaging course, evaluate its retrieval accuracy, latency characteristics, and user experience, and show that ALIVE provides accurate, content-aware, and engaging real-time support. ALIVE illustrates how multimodal AI-when combined with content-aware retrieval and local deployment-can significantly enhance the pedagogical value of recorded lectures, offering an extensible pathway toward next-generation interactive learning environments.

Method: 1) B/C/D-Scan three-view joint analysis with FDTD simulation validation; 2) DCO-YOLO framework integrating DySample, CGLU, and OutlookAttention into YOLOv11; 3) 3D-DIoU spatial feature matching with geometric constraints and center distance penalty; 4) Three-view fusion strategy.

Result: Achieves 96.2% accuracy, 93.3% recall, and 96.7% mAP in complex multi-pipeline scenarios, outperforming baseline by 2.0%, 2.1%, and 0.9% respectively. Ablation studies validate module effectiveness, and Grad-CAM++ shows improved focus on pipeline geometry.

Conclusion: Integrates deep learning optimization with 3D GPR physical characteristics, providing efficient and reliable technical framework for intelligent underground pipeline recognition and localization.

Abstract: To address the issues of weak correlation between multi-view features, low recognition accuracy of small-scale targets, and insufficient robustness in complex scenarios in underground pipeline detection using 3D GPR, this paper proposes a 3D pipeline intelligent detection framework. First, based on a B/C/D-Scan three-view joint analysis strategy, a three-dimensional pipeline three-view feature evaluation method is established by cross-validating forward simulation results obtained using FDTD methods with actual measurement data. Second, the DCO-YOLO framework is proposed, which integrates DySample, CGLU, and OutlookAttention cross-dimensional correlation mechanisms into the original YOLOv11 algorithm, significantly improving the small-scale pipeline edge feature extraction capability. Furthermore, a 3D-DIoU spatial feature matching algorithm is proposed, which integrates three-dimensional geometric constraints and center distance penalty terms to achieve automated association of multi-view annotations. The three-view fusion strategy resolves inherent ambiguities in single-view detection. Experiments based on real urban underground pipeline data show that the proposed method achieves accuracy, recall, and mean average precision of 96.2%, 93.3%, and 96.7%, respectively, in complex multi-pipeline scenarios, which are 2.0%, 2.1%, and 0.9% higher than the baseline model. Ablation experiments validated the synergistic optimization effect of the dynamic feature enhancement module and Grad-CAM++ heatmap visualization demonstrated that the improved model significantly enhanced its ability to focus on pipeline geometric features. This study integrates deep learning optimization strategies with the physical characteristics of 3D GPR, offering an efficient and reliable novel technical framework for the intelligent recognition and localization of underground pipelines.

Method: DRI keeps VFM frozen, uses lightweight Offset Encoder to extract domain-specific representations, adaptively transforms them via Modulator using contextual information, and injects them into intermediate layers via additive fusion without altering pre-trained weights.

Result: Achieves SOTA performance with minimal parameters: 57.9% and 60.5% mAP on HOSS-ReID dataset using only 1.54M and 7.05M parameters respectively.

Conclusion: DRI effectively bridges modality gaps in cross-modality ship re-ID by optimizing in feature space rather than weight space, preserving general knowledge while adapting to downstream tasks with minimal trainable parameters.

Abstract: Cross-Modality Ship Re-Identification (CMS Re-ID) is critical for achieving all-day and all-weather maritime target tracking, yet it is fundamentally challenged by significant modality discrepancies. Mainstream solutions typically rely on explicit modality alignment strategies; however, this paradigm heavily depends on constructing large-scale paired datasets for pre-training. To address this, grounded in the Platonic Representation Hypothesis, we explore the potential of Vision Foundation Models (VFMs) in bridging modality gaps. Recognizing the suboptimal performance of existing generic Parameter-Efficient Fine-Tuning (PEFT) methods that operate within the weight space, particularly on limited-capacity models, we shift the optimization perspective to the feature space and propose a novel PEFT strategy termed Domain Representation Injection (DRI). Specifically, while keeping the VFM fully frozen to maximize the preservation of general knowledge, we design a lightweight, learnable Offset Encoder to extract domain-specific representations rich in modality and identity attributes from raw inputs. Guided by the contextual information of intermediate features at different layers, a Modulator adaptively transforms these representations. Subsequently, they are injected into the intermediate layers via additive fusion, dynamically reshaping the feature distribution to adapt to the downstream task without altering the VFM’s pre-trained weights. Extensive experimental results demonstrate the superiority of our method, achieving State-of-the-Art (SOTA) performance with minimal trainable parameters. For instance, on the HOSS-ReID dataset, we attain 57.9% and 60.5% mAP using only 1.54M and 7.05M parameters, respectively. The code is available at https://github.com/TingfengXian/DRI.

Method: Proposed DGSAN with: 1) Global-Local Feature Encoder to capture local, global, and fused nodule characteristics; 2) Dual-Graph Construction organizing multimodal features into inter-modal and intra-modal graphs; 3) Hierarchical Cross-Modal Graph Fusion Module for refined feature integration. Also created NLST-cmst multimodal dataset.

Result: Extensive experiments on NLST-cmst and CSTL-derived datasets show DGSAN significantly outperforms state-of-the-art methods in pulmonary nodule classification with exceptional computational efficiency.

Conclusion: The proposed DGSAN framework effectively addresses limitations in multimodal fusion for lung nodule analysis, demonstrating superior performance and efficiency through innovative graph-based spatiotemporal attention mechanisms.

Abstract: Lung cancer continues to be the leading cause of cancer-related deaths globally. Early detection and diagnosis of pulmonary nodules are essential for improving patient survival rates. Although previous research has integrated multimodal and multi-temporal information, outperforming single modality and single time point, the fusion methods are limited to inefficient vector concatenation and simple mutual attention, highlighting the need for more effective multimodal information fusion. To address these challenges, we introduce a Dual-Graph Spatiotemporal Attention Network, which leverages temporal variations and multimodal data to enhance the accuracy of predictions. Our methodology involves developing a Global-Local Feature Encoder to better capture the local, global, and fused characteristics of pulmonary nodules. Additionally, a Dual-Graph Construction method organizes multimodal features into inter-modal and intra-modal graphs. Furthermore, a Hierarchical Cross-Modal Graph Fusion Module is introduced to refine feature integration. We also compiled a novel multimodal dataset named the NLST-cmst dataset as a comprehensive source of support for related research. Our extensive experiments, conducted on both the NLST-cmst and curated CSTL-derived datasets, demonstrate that our DGSAN significantly outperforms state-of-the-art methods in classifying pulmonary nodules with exceptional computational efficiency.

Method: 1) Created comprehensive benchmark with 1M+ compressed images using various codecs and tasks; 2) Analyzed performance gaps into information loss vs. generalization failure; 3) Developed universal VLM adaptor to enhance performance on compressed images.

Result: Benchmark reveals significant performance degradation with low-bitrate images; analysis shows generalization gap is the main issue (not information loss); universal adaptor improves VLM performance by 10-30% across different codecs and bitrates.

Conclusion: The benchmark provides valuable insights into VLM performance on compressed images, and the proposed universal adaptor effectively bridges the gap between VLMs and compressed images, making VLMs more practical for real-world applications.

Abstract: With the rapid development of Vision-Language Models (VLMs) and the growing demand for their applications, efficient compression of the image inputs has become increasingly important. Existing VLMs predominantly digest and understand high-bitrate compressed images, while their ability to interpret low-bitrate compressed images has yet to be explored by far. In this paper, we introduce the first comprehensive benchmark to evaluate the ability of VLM against compressed images, varying existing widely used image codecs and diverse set of tasks, encompassing over one million compressed images in our benchmark. Next, we analyse the source of performance gap, by categorising the gap from a) the information loss during compression and b) generalisation failure of VLM. We visualize these gaps with concrete examples and identify that for compressed images, only the generalization gap can be mitigated. Finally, we propose a universal VLM adaptor to enhance model performance on images compressed by existing codecs. Consequently, we demonstrate that a single adaptor can improve VLM performance across images with varying codecs and bitrates by 10%-30%. We believe that our benchmark and enhancement method provide valuable insights and contribute toward bridging the gap between VLMs and compressed images.

Method: Three-stage pipeline: 1) Places compact set of panoramic viewpoints considering scene layout/geometry, 2) Grounds text query on each panoramic rendering using a pretrained 2D VLM, 3) Fuses per-view predictions into single 3D bounding box via lifting. Uses panoramic renderings augmented with 3D semantic/geometric features as intermediate representation.

Result: Achieves state-of-the-art results on ScanRefer and Nr3D benchmarks. Demonstrates superior generalization to unseen 3D datasets and text rephrasings compared to traditional supervised models.

Conclusion: PanoGrounder successfully bridges 2D VLMs with 3D scene understanding through panoramic representations, enabling strong vision-language reasoning while maintaining 3D geometric awareness, leading to improved generalization in 3D visual grounding tasks.

Abstract: 3D Visual Grounding (3DVG) is a critical bridge from vision-language perception to robotics, requiring both language understanding and 3D scene reasoning. Traditional supervised models leverage explicit 3D geometry but exhibit limited generalization, owing to the scarcity of 3D vision-language datasets and the limited reasoning capabilities compared to modern vision-language models (VLMs). We propose PanoGrounder, a generalizable 3DVG framework that couples multi-modal panoramic representation with pretrained 2D VLMs for strong vision-language reasoning. Panoramic renderings, augmented with 3D semantic and geometric features, serve as an intermediate representation between 2D and 3D, and offer two major benefits: (i) they can be directly fed to VLMs with minimal adaptation and (ii) they retain long-range object-to-object relations thanks to their 360-degree field of view. We devise a three-stage pipeline that places a compact set of panoramic viewpoints considering the scene layout and geometry, grounds a text query on each panoramic rendering with a VLM, and fuses per-view predictions into a single 3D bounding box via lifting. Our approach achieves state-of-the-art results on ScanRefer and Nr3D, and demonstrates superior generalization to unseen 3D datasets and text rephrasings.

Method: Proposes SMC-Mamba framework with: 1) Modality-Agnostic Feature Enhancement (MAFE) module using adaptive gating and spatial-channel/frequency-rotational scanning; 2) Multiplex Consensus Cross-modal Mamba (MCCM) module for dynamic expert collaboration and cross-modal feature interaction; 3) Bi-level Self-supervised Contrastive Learning Loss (BSCL) to preserve high-frequency information without computational overhead.

Result: Extensive experiments show the approach outperforms state-of-the-art image fusion algorithms in infrared-visible, medical, multi-focus, and multi-exposure fusion tasks, as well as downstream visual tasks.

Conclusion: SMC-Mamba provides an effective general image fusion framework that enhances performance across diverse fusion tasks and downstream applications through innovative self-supervised learning and cross-modal integration techniques.

Abstract: Image fusion integrates complementary information from different modalities to generate high-quality fused images, thereby enhancing downstream tasks such as object detection and semantic segmentation. Unlike task-specific techniques that primarily focus on consolidating inter-modal information, general image fusion needs to address a wide range of tasks while improving performance without increasing complexity. To achieve this, we propose SMC-Mamba, a Self-supervised Multiplex Consensus Mamba framework for general image fusion. Specifically, the Modality-Agnostic Feature Enhancement (MAFE) module preserves fine details through adaptive gating and enhances global representations via spatial-channel and frequency-rotational scanning. The Multiplex Consensus Cross-modal Mamba (MCCM) module enables dynamic collaboration among experts, reaching a consensus to efficiently integrate complementary information from multiple modalities. The cross-modal scanning within MCCM further strengthens feature interactions across modalities, facilitating seamless integration of critical information from both sources. Additionally, we introduce a Bi-level Self-supervised Contrastive Learning Loss (BSCL), which preserves high-frequency information without increasing computational overhead while simultaneously boosting performance in downstream tasks. Extensive experiments demonstrate that our approach outperforms state-of-the-art (SOTA) image fusion algorithms in tasks such as infrared-visible, medical, multi-focus, and multi-exposure fusion, as well as downstream visual tasks.

Method: Proposes Quantile Rendering (Q-Render) that sparsely samples only dominant Gaussians along each ray instead of all intersecting ones. Also introduces Gaussian Splatting Network (GS-Net), a generalizable 3D neural network that predicts Gaussian features.

Result: Outperforms state-of-the-art methods on ScanNet and LeRF benchmarks while enabling real-time rendering with ~43.7x speedup on 512-D feature maps compared to conventional approaches.

Conclusion: Q-Render provides an efficient solution for high-dimensional feature rendering in 3D open-vocabulary segmentation, maintaining high fidelity while achieving significant speed improvements, making real-time applications feasible.

Abstract: Recent advancements in computer vision have successfully extended Open-vocabulary segmentation (OVS) to the 3D domain by leveraging 3D Gaussian Splatting (3D-GS). Despite this progress, efficiently rendering the high-dimensional features required for open-vocabulary queries poses a significant challenge. Existing methods employ codebooks or feature compression, causing information loss, thereby degrading segmentation quality. To address this limitation, we introduce Quantile Rendering (Q-Render), a novel rendering strategy for 3D Gaussians that efficiently handles high-dimensional features while maintaining high fidelity. Unlike conventional volume rendering, which densely samples all 3D Gaussians intersecting each ray, Q-Render sparsely samples only those with dominant influence along the ray. By integrating Q-Render into a generalizable 3D neural network, we also propose Gaussian Splatting Network (GS-Net), which predicts Gaussian features in a generalizable manner. Extensive experiments on ScanNet and LeRF demonstrate that our framework outperforms state-of-the-art methods, while enabling real-time rendering with an approximate ~43.7x speedup on 512-D feature maps. Code will be made publicly available.

Method: Collaborative Multi-Agent Reasoning Framework that explicitly decouples Semantic Planning from Visual Synthesis. Includes: (1) self-correcting Verification Agent using Chain-of-Thought reasoning to correct segmentation and identify occluders, and (2) Diverse Hypothesis Generator for multiple plausible semantic interpretations. Also introduces MAC-Score evaluation metric.

Result: Framework significantly outperforms state-of-the-art methods across multiple datasets. The MAC-Score metric is validated against human judgment and ground truth, establishing a robust standard for assessing structural completeness and semantic consistency.

Conclusion: The proposed collaborative multi-agent reasoning approach effectively addresses limitations of prior methods by enabling explicit semantic planning before visual synthesis, leading to more coherent and accurate amodal completion results with better semantic consistency.

Abstract: Amodal completion, the task of inferring invisible object parts, faces significant challenges in maintaining semantic consistency and structural integrity. Prior progressive approaches are inherently limited by inference instability and error accumulation. To tackle these limitations, we present a Collaborative Multi-Agent Reasoning Framework that explicitly decouples Semantic Planning from Visual Synthesis. By employing specialized agents for upfront reasoning, our method generates a structured, explicit plan before pixel generation, enabling visually and semantically coherent single-pass synthesis. We integrate this framework with two critical mechanisms: (1) a self-correcting Verification Agent that employs Chain-of-Thought reasoning to rectify visible region segmentation and identify residual occluders strictly within the Semantic Planning phase, and (2) a Diverse Hypothesis Generator that addresses the ambiguity of invisible regions by offering diverse, plausible semantic interpretations, surpassing the limited pixel-level variations of standard random seed sampling. Furthermore, addressing the limitations of traditional metrics in assessing inferred invisible content, we introduce the MAC-Score (MLLM Amodal Completion Score), a novel human-aligned evaluation metric. Validated against human judgment and ground truth, these metrics establish a robust standard for assessing structural completeness and semantic consistency with visible context. Extensive experiments demonstrate that our framework significantly outperforms state-of-the-art methods across multiple datasets. Our project is available at: https://fanhongxing.github.io/remac-page.

Method: Real-centric Envelope Modeling (REM) shifts detection from learning generator artifacts to modeling robust real image distributions. It introduces feature-level perturbations in self-reconstruction to generate near-real samples, and uses an envelope estimator with cross-domain consistency to learn a boundary enclosing the real image manifold.

Result: REM achieves average 7.5% improvement over state-of-the-art methods across eight benchmark evaluations. It maintains exceptional generalization on the severely degraded RealChain benchmark, establishing solid foundation for synthetic image detection under real-world conditions.

Conclusion: REM provides a new paradigm for synthetic image detection that is more robust to real-world degradations and evolving generative models by focusing on modeling real image distributions rather than generator artifacts.

Abstract: The rapid progress of generative models has intensified the need for reliable and robust detection under real-world conditions. However, existing detectors often overfit to generator-specific artifacts and remain highly sensitive to real-world degradations. As generative architectures evolve and images undergo multi-round cross-platform sharing and post-processing (chain degradations), these artifact cues become obsolete and harder to detect. To address this, we propose Real-centric Envelope Modeling (REM), a new paradigm that shifts detection from learning generator artifacts to modeling the robust distribution of real images. REM introduces feature-level perturbations in self-reconstruction to generate near-real samples, and employs an envelope estimator with cross-domain consistency to learn a boundary enclosing the real image manifold. We further build RealChain, a comprehensive benchmark covering both open-source and commercial generators with simulated real-world degradation. Across eight benchmark evaluations, REM achieves an average improvement of 7.5% over state-of-the-art methods, and notably maintains exceptional generalization on the severely degraded RealChain benchmark, establishing a solid foundation for synthetic image detection under real-world conditions. The code and the RealChain benchmark will be made publicly available upon acceptance of the paper.

Method: Uses map-guided LLM agent with road structures and CCTV placement as spatial documents, transforms vehicle positions to world coordinates, combines spatial info with movement patterns, and performs beam search at intersections to predict next CCTV location.

Result: Experimental results in CARLA simulator show SPOT accurately predicts next appearing CCTV in blind spots and maintains continuous vehicle trajectories better than existing techniques.

Conclusion: SPOT effectively addresses blind spot limitations in multi-CCTV tracking through spatial reasoning and map guidance, enabling reliable continuous vehicle tracking without prior training.

Abstract: CCTV-based vehicle tracking systems face structural limitations in continuously connecting the trajectories of the same vehicle across multiple camera environments. In particular, blind spots occur due to the intervals between CCTVs and limited Fields of View (FOV), which leads to object ID switching and trajectory loss, thereby reducing the reliability of real-time path prediction. This paper proposes SPOT (Spatial Prediction Over Trajectories), a map-guided LLM agent capable of tracking vehicles even in blind spots of multi-CCTV environments without prior training. The proposed method represents road structures (Waypoints) and CCTV placement information as documents based on 2D spatial coordinates and organizes them through chunking techniques to enable real-time querying and inference. Furthermore, it transforms the vehicle’s position into the actual world coordinate system using the relative position and FOV information of objects observed in CCTV images. By combining map spatial information with the vehicle’s moving direction, speed, and driving patterns, a beam search is performed at the intersection level to derive candidate CCTV locations where the vehicle is most likely to enter after the blind spot. Experimental results based on the CARLA simulator in a virtual city environment confirmed that the proposed method accurately predicts the next appearing CCTV even in blind spot sections, maintaining continuous vehicle trajectories more effectively than existing techniques.

Method: Hybrid grid framework combining sparse grid (geometric priors) with implicit dense grid (scene enrichment), using VDB structure with submap organization. Includes distillation-based overlap alignment for consistency and dynamic removal module for robustness.

Result: Superior mapping quality while overcoming efficiency limitations of voxel-guided methods, outperforming state-of-the-art mapping methods in extensive experiments.

Conclusion: XGrid-Mapping successfully addresses the efficiency-quality trade-off in neural LiDAR mapping, enabling real-time large-scale incremental mapping with improved geometric utilization and consistency.

Abstract: Large-scale incremental mapping is fundamental to the development of robust and reliable autonomous systems, as it underpins incremental environmental understanding with sequential inputs for navigation and decision-making. LiDAR is widely used for this purpose due to its accuracy and robustness. Recently, neural LiDAR mapping has shown impressive performance; however, most approaches rely on dense implicit representations and underutilize geometric structure, while existing voxel-guided methods struggle to achieve real-time performance. To address these challenges, we propose XGrid-Mapping, a hybrid grid framework that jointly exploits explicit and implicit representations for efficient neural LiDAR mapping. Specifically, the strategy combines a sparse grid, providing geometric priors and structural guidance, with an implicit dense grid that enriches scene representation. By coupling the VDB structure with a submap-based organization, the framework reduces computational load and enables efficient incremental mapping on a large scale. To mitigate discontinuities across submaps, we introduce a distillation-based overlap alignment strategy, in which preceding submaps supervise subsequent ones to ensure consistency in overlapping regions. To further enhance robustness and sampling efficiency, we incorporate a dynamic removal module. Extensive experiments show that our approach delivers superior mapping quality while overcoming the efficiency limitations of voxel-guided methods, thereby outperforming existing state-of-the-art mapping methods.

Method: Proposed pipeline: 1) Train diffusion model on abundant normal X-rays, 2) Use pre-trained model to inpaint head lesions onto diseased X-rays to create augmented tail class data, 3) Integrate Large Language Model Knowledge Guidance (LKG) module with Progressive Incremental Learning (PIL) strategy to stabilize fine-tuning.

Result: Comprehensive evaluations on public lung datasets MIMIC and CheXpert demonstrate the method sets new benchmark in performance for long-tailed pulmonary anomaly detection.

Conclusion: The proposed approach effectively addresses data scarcity for rare lung lesions through innovative synthesis using normal X-rays and advanced stabilization techniques, achieving state-of-the-art diagnostic performance.

Abstract: Long-tailed pulmonary anomalies in chest radiography present formidable diagnostic challenges. Despite the recent strides in diffusion-based methods for enhancing the representation of tailed lesions, the paucity of rare lesion exemplars curtails the generative capabilities of these approaches, thereby leaving the diagnostic precision less than optimal. In this paper, we propose a novel data synthesis pipeline designed to augment tail lesions utilizing a copious supply of conventional normal X-rays. Specifically, a sufficient quantity of normal samples is amassed to train a diffusion model capable of generating normal X-ray images. This pre-trained diffusion model is subsequently utilized to inpaint the head lesions present in the diseased X-rays, thereby preserving the tail classes as augmented training data. Additionally, we propose the integration of a Large Language Model Knowledge Guidance (LKG) module alongside a Progressive Incremental Learning (PIL) strategy to stabilize the inpainting fine-tuning process. Comprehensive evaluations conducted on the public lung datasets MIMIC and CheXpert demonstrate that the proposed method sets a new benchmark in performance.

Method: Two-stage flow-matching: first learns straight-path flow from sparse to dense targets, then refines with noise-perturbed samples. Includes data-driven adaptive time scheduler for efficient sampling, on-manifold constraints during sampling, and recurrent interface network (RIN) for hierarchical feature interactions.

Result: Extensive experiments on synthetic benchmarks and real-world scans show PUFM++ sets new state-of-the-art in point cloud upsampling, delivering superior visual fidelity and quantitative accuracy across various tasks.

Conclusion: PUFM++ presents an enhanced flow-matching framework that advances point cloud upsampling through improved geometric fidelity, robustness, and downstream task consistency, with publicly available code and pretrained models.

Abstract: Recent advances in generative modeling have demonstrated strong promise for high-quality point cloud upsampling. In this work, we present PUFM++, an enhanced flow-matching framework for reconstructing dense and accurate point clouds from sparse, noisy, and partial observations. PUFM++ improves flow matching along three key axes: (i) geometric fidelity, (ii) robustness to imperfect input, and (iii) consistency with downstream surface-based tasks. We introduce a two-stage flow-matching strategy that first learns a direct, straight-path flow from sparse inputs to dense targets, and then refines it using noise-perturbed samples to approximate the terminal marginal distribution better. To accelerate and stabilize inference, we propose a data-driven adaptive time scheduler that improves sampling efficiency based on interpolation behavior. We further impose on-manifold constraints during sampling to ensure that generated points remain aligned with the underlying surface. Finally, we incorporate a recurrent interface network~(RIN) to strengthen hierarchical feature interactions and boost reconstruction quality. Extensive experiments on synthetic benchmarks and real-world scans show that PUFM++ sets a new state of the art in point cloud upsampling, delivering superior visual fidelity and quantitative accuracy across a wide range of tasks. Code and pretrained models are publicly available at https://github.com/Holmes-Alan/Enhanced_PUFM.

Method: Feed-forward framework using alternating attention across views to capture intra-view lighting interactions and inter-view material consistency. Includes consistency-based finetuning strategy using unlabeled real-world videos to enhance robustness.

Result: Achieves state-of-the-art performance in multi-view consistency, material/normal estimation quality, and generalization to real-world imagery on benchmark datasets.

Conclusion: The proposed framework enables coherent scene-level reasoning in a single forward pass, overcoming limitations of both single-view and optimization-based multi-view approaches while improving real-world generalization.

Abstract: Multi-view inverse rendering aims to recover geometry, materials, and illumination consistently across multiple viewpoints. When applied to multi-view images, existing single-view approaches often ignore cross-view relationships, leading to inconsistent results. In contrast, multi-view optimization methods rely on slow differentiable rendering and per-scene refinement, making them computationally expensive and hard to scale. To address these limitations, we introduce a feed-forward multi-view inverse rendering framework that directly predicts spatially varying albedo, metallic, roughness, diffuse shading, and surface normals from sequences of RGB images. By alternating attention across views, our model captures both intra-view long-range lighting interactions and inter-view material consistency, enabling coherent scene-level reasoning within a single forward pass. Due to the scarcity of real-world training data, models trained on existing synthetic datasets often struggle to generalize to real-world scenes. To overcome this limitation, we propose a consistency-based finetuning strategy that leverages unlabeled real-world videos to enhance both multi-view coherence and robustness under in-the-wild conditions. Extensive experiments on benchmark datasets demonstrate that our method achieves state-of-the-art performance in terms of multi-view consistency, material and normal estimation quality, and generalization to real-world imagery.

Method: Proposes NExT-Vid with two key components: 1) context-isolated autoregressive predictor to decouple semantic representation from target decoding, and 2) conditioned flow-matching decoder to enhance generation quality and diversity. Uses masked next-frame prediction for joint image-video modeling.

Result: Extensive experiments on large-scale pretrained models show NExT-Vid consistently outperforms previous generative pretraining methods for visual representation learning via attentive probing in downstream classification tasks.

Conclusion: NExT-Vid successfully addresses limitations of existing autoregressive visual pretraining methods and demonstrates strong performance in visual representation learning through its novel architecture combining context isolation and flow-matching techniques.

Abstract: Recent advances in pretraining general foundation models have significantly improved performance across diverse downstream tasks. While autoregressive (AR) generative models like GPT have revolutionized NLP, most visual generative pretraining methods still rely on BERT-style masked modeling, which often disregards the temporal information essential for video analysis. The few existing autoregressive visual pretraining methods suffer from issues such as inaccurate semantic localization and poor generation quality, leading to poor semantics. In this work, we propose NExT-Vid, a novel autoregressive visual generative pretraining framework that utilizes masked next-frame prediction to jointly model images and videos. NExT-Vid introduces a context-isolated autoregressive predictor to decouple semantic representation from target decoding, and a conditioned flow-matching decoder to enhance generation quality and diversity. Through context-isolated flow-matching pretraining, our approach achieves strong representations. Extensive experiments on large-scale pretrained models demonstrate that our proposed method consistently outperforms previous generative pretraining methods for visual representation learning via attentive probing in downstream classification.

Method: Granular-ball Guided Masking (GBGM) uses Granular-ball Computing (GBC) to guide a coarse-to-fine hierarchical masking process. It adaptively preserves semantically rich, structurally important regions while suppressing redundant areas.

Result: Extensive experiments on multiple benchmarks show consistent improvements in classification accuracy and masked image reconstruction. The method works with both CNNs and Vision Transformers.

Conclusion: GBGM provides an effective, model-agnostic structure-aware data augmentation paradigm that enhances model robustness and generalizability without requiring architectural modifications.

Abstract: Deep learning models have achieved remarkable success in computer vision, but they still rely heavily on large-scale labeled data and tend to overfit when data are limited or distributions shift. Data augmentation, particularly mask-based information dropping, can enhance robustness by forcing models to explore complementary cues; however, existing approaches often lack structural awareness and may discard essential semantics. We propose Granular-ball Guided Masking (GBGM), a structure-aware augmentation strategy guided by Granular-ball Computing (GBC). GBGM adaptively preserves semantically rich, structurally important regions while suppressing redundant areas through a coarse-to-fine hierarchical masking process, producing augmentations that are both representative and discriminative. Extensive experiments on multiple benchmarks demonstrate consistent improvements in classification accuracy and masked image reconstruction, confirming the effectiveness and broad applicability of the proposed method. Simple and model-agnostic, it integrates seamlessly into CNNs and Vision Transformers and provides a new paradigm for structure-aware data augmentation.

Method: Integrates Mamba (linear time-series module) into pretrained Stable Diffusion models to replace temporal attention layers, uses low-rank approximation for query/key weight matrices, and employs weighted averaging during training to update attention scores.

Result: Demonstrates promising results in editing various attributes, subjects, and locations in real-world videos with improved temporal consistency and reduced computational burden.

Conclusion: FluencyVE provides an effective one-shot video editing approach that preserves the generative power of text-to-image models while addressing temporal inconsistency and computational efficiency challenges.

Abstract: Large-scale text-to-image diffusion models have achieved unprecedented success in image generation and editing. However, extending this success to video editing remains challenging. Recent video editing efforts have adapted pretrained text-to-image models by adding temporal attention mechanisms to handle video tasks. Unfortunately, these methods continue to suffer from temporal inconsistency issues and high computational overheads. In this study, we propose FluencyVE, which is a simple yet effective one-shot video editing approach. FluencyVE integrates the linear time-series module, Mamba, into a video editing model based on pretrained Stable Diffusion models, replacing the temporal attention layer. This enables global frame-level attention while reducing the computational costs. In addition, we employ low-rank approximation matrices to replace the query and key weight matrices in the causal attention, and use a weighted averaging technique during training to update the attention scores. This approach significantly preserves the generative power of the text-to-image model while effectively reducing the computational burden. Experiments and analyses demonstrate promising results in editing various attributes, subjects, and locations in real-world videos.

Method: Proposes an efficient video defense framework that protects portrait videos by perturbing the 3D information acquisition process. Key innovations include: (1) a similarity-guided parameter sharing mechanism for computational efficiency, and (2) a multi-scale dual-domain attention module to jointly optimize spatial-frequency perturbations.

Result: Extensive experiments show strong defense capability with 47x acceleration over the fastest baseline while maintaining high fidelity. The framework remains robust against scaling operations and state-of-the-art purification attacks, with design choices validated through ablation studies.

Conclusion: The proposed framework provides an efficient and effective solution for protecting portrait videos against 3D-field talking face generation attacks, addressing privacy concerns while maintaining video quality and computational efficiency.

Abstract: State-of-the-art 3D-field video-referenced Talking Face Generation (TFG) methods synthesize high-fidelity personalized talking-face videos in real time by modeling 3D geometry and appearance from reference portrait video. This capability raises significant privacy concerns regarding malicious misuse of personal portraits. However, no efficient defense framework exists to protect such videos against 3D-field TFG methods. While image-based defenses could apply per-frame 2D perturbations, they incur prohibitive computational costs, severe video quality degradation, failing to disrupt 3D information for video protection. To address this, we propose a novel and efficient video defense framework against 3D-field TFG methods, which protects portrait video by perturbing the 3D information acquisition process while maintain high-fidelity video quality. Specifically, our method introduces: (1) a similarity-guided parameter sharing mechanism for computational efficiency, and (2) a multi-scale dual-domain attention module to jointly optimize spatial-frequency perturbations. Extensive experiments demonstrate that our proposed framework exhibits strong defense capability and achieves a 47x acceleration over the fastest baseline while maintaining high fidelity. Moreover, it remains robust against scaling operations and state-of-the-art purification attacks, and the effectiveness of our design choices is further validated through ablation studies. Our project is available at https://github.com/Richen7418/VDF.

Method: Latent diffusion model for high-quality visible face generation from thermal inputs, multi-attribute classifier to preserve identity features, and Self-attn Mamba module for cross-modal feature modeling and faster inference.

Result: Achieves state-of-the-art performance on two benchmark datasets, with superior image quality and identity preservation compared to existing methods.

Conclusion: The proposed approach effectively addresses domain shift in infrared face recognition by generating high-quality visible images while preserving identity features, advancing HFR technology.

Abstract: Modern surveillance systems increasingly rely on multi-wavelength sensors and deep neural networks to recognize faces in infrared images captured at night. However, most facial recognition models are trained on visible light datasets, leading to substantial performance degradation on infrared inputs due to significant domain shifts. Early feature-based methods for infrared face recognition proved ineffective, prompting researchers to adopt generative approaches that convert infrared images into visible light images for improved recognition. This paradigm, known as Heterogeneous Face Recognition (HFR), faces challenges such as model and modality discrepancies, leading to distortion and feature loss in generated images. To address these limitations, this paper introduces a novel latent diffusion-based model designed to generate high-quality visible face images from thermal inputs while preserving critical identity features. A multi-attribute classifier is incorporated to extract key facial attributes from visible images, mitigating feature loss during infrared-to-visible image restoration. Additionally, we propose the Self-attn Mamba module, which enhances global modeling of cross-modal features and significantly improves inference speed. Experimental results on two benchmark datasets demonstrate the superiority of our approach, achieving state-of-the-art performance in both image quality and identity preservation.

Method: Next-Scale Prediction (NSP) constructs cross-scale training pairs where blind-spot networks take low-resolution, fully decorrelated sub-images as input to predict high-resolution targets that retain fine details. This decouples noise decorrelation from detail preservation.

Result: NSP achieves state-of-the-art self-supervised denoising performance on real-world benchmarks, significantly alleviating the conflict between noise decorrelation and detail preservation. As a by-product, it naturally supports super-resolution of noisy images without retraining.

Conclusion: NSP provides an effective solution to the long-standing trade-off in self-supervised denoising by separating noise decorrelation from detail preservation through cross-scale prediction, enabling superior performance while naturally supporting super-resolution tasks.

Abstract: Self-supervised real-world image denoising remains a fundamental challenge, arising from the antagonistic trade-off between decorrelating spatially structured noise and preserving high-frequency details. Existing blind-spot network (BSN) methods rely on pixel-shuffle downsampling (PD) to decorrelate noise, but aggressive downsampling fragments fine structures, while milder downsampling fails to remove correlated noise. To address this, we introduce Next-Scale Prediction (NSP), a novel self-supervised paradigm that decouples noise decorrelation from detail preservation. NSP constructs cross-scale training pairs, where BSN takes low-resolution, fully decorrelated sub-images as input to predict high-resolution targets that retain fine details. As a by-product, NSP naturally supports super-resolution of noisy images without retraining or modification. Extensive experiments demonstrate that NSP achieves state-of-the-art self-supervised denoising performance on real-world benchmarks, significantly alleviating the long-standing conflict between noise decorrelation and detail preservation.

Method: Created KOREATECH-CGH dataset with 6,000 RGB-D image/hologram pairs across resolutions from 256×256 to 2048×2048. Introduced amplitude projection technique that replaces amplitude components of hologram wavefields at each depth layer while preserving phase.

Result: Achieved 27.01 dB PSNR and 0.87 SSIM, surpassing recent optimized silhouette-masking layer method by 2.03 dB and 0.04 SSIM. Dataset validated through hologram generation and super-resolution experiments with state-of-the-art ML models.

Conclusion: KOREATECH-CGH dataset and amplitude projection technique provide valuable resources for training and evaluating next-generation ML-CGH systems, addressing the data scarcity problem in the field.

Abstract: Machine learning-based computer-generated holography (ML-CGH) has advanced rapidly in recent years, yet progress is constrained by the limited availability of high-quality, large-scale hologram datasets. To address this, we present KOREATECH-CGH, a publicly available dataset comprising 6,000 pairs of RGB-D images and complex holograms across resolutions ranging from 256256 to 20482048, with depth ranges extending to the theoretical limits of the angular spectrum method for wide 3D scene coverage. To improve hologram quality at large depth ranges, we introduce amplitude projection, a post-processing technique that replaces amplitude components of hologram wavefields at each depth layer while preserving phase. This approach enhances reconstruction fidelity, achieving 27.01 dB PSNR and 0.87 SSIM, surpassing a recent optimized silhouette-masking layer-based method by 2.03 dB and 0.04 SSIM, respectively. We further validate the utility of KOREATECH-CGH through experiments on hologram generation and super-resolution using state-of-the-art ML models, confirming its applicability for training and evaluating next-generation ML-CGH systems.

Method: Introduces a reweighted logarithmic norm as a nonconvex surrogate for rank function, then solves the optimization problem using alternating direction method of multipliers (ADMM).

Result: Experimental results on image inpainting show superior performance compared to state-of-the-art LRMC approaches in both visual quality and quantitative metrics.

Conclusion: The proposed reweighted logarithmic norm provides a closer approximation to the rank function than existing alternatives, leading to better performance in low-rank matrix completion tasks.

Abstract: Low-rank matrix completion (LRMC) has demonstrated remarkable success in a wide range of applications. To address the NP-hard nature of the rank minimization problem, the nuclear norm is commonly used as a convex and computationally tractable surrogate for the rank function. However, this approach often yields suboptimal solutions due to the excessive shrinkage of singular values. In this letter, we propose a novel reweighted logarithmic norm as a more effective nonconvex surrogate, which provides a closer approximation than many existing alternatives. We efficiently solve the resulting optimization problem by employing the alternating direction method of multipliers (ADMM). Experimental results on image inpainting demonstrate that the proposed method achieves superior performance compared to state-of-the-art LRMC approaches, both in terms of visual quality and quantitative metrics.

Method: Uses 2D-3D hybrid feature extraction to detect corners and edges from events and object models. Optical flow of corners is found by maximizing event-associated probability, then establishes corner-edge correlation. 6DoF pose is optimized by minimizing distances between corners and edges.

Result: Outperforms event-based state-of-the-art methods in both accuracy and robustness on simulated and real event data.

Conclusion: The optical flow-guided approach with event cameras provides an effective solution for robust 6DoF object pose tracking, overcoming limitations of traditional cameras.

Abstract: Object pose tracking is one of the pivotal technologies in multimedia, attracting ever-growing attention in recent years. Existing methods employing traditional cameras encounter numerous challenges such as motion blur, sensor noise, partial occlusion, and changing lighting conditions. The emerging bio-inspired sensors, particularly event cameras, possess advantages such as high dynamic range and low latency, which hold the potential to address the aforementioned challenges. In this work, we present an optical flow-guided 6DoF object pose tracking method with an event camera. A 2D-3D hybrid feature extraction strategy is firstly utilized to detect corners and edges from events and object models, which characterizes object motion precisely. Then, we search for the optical flow of corners by maximizing the event-associated probability within a spatio-temporal window, and establish the correlation between corners and edges guided by optical flow. Furthermore, by minimizing the distances between corners and edges, the 6DoF object pose is iteratively optimized to achieve continuous pose tracking. Experimental results of both simulated and real events demonstrate that our methods outperform event-based state-of-the-art methods in terms of both accuracy and robustness.

Method: DexAvatar uses a framework guided by learned 3D hand and body priors to reconstruct bio-mechanically accurate fine-grained hand articulations and body movements from in-the-wild monocular sign language videos.

Result: Achieves 35.11% improvement in body and hand pose estimation compared to state-of-the-art on the SGNify motion capture dataset, the only available benchmark for this task.

Conclusion: DexAvatar provides a significant advancement in reconstructing accurate 3D poses from sign language videos, addressing critical limitations in current pose estimation methods for sign language generation applications.

Abstract: The trend in sign language generation is centered around data-driven generative methods that require vast amounts of precise 2D and 3D human pose data to achieve an acceptable generation quality. However, currently, most sign language datasets are video-based and limited to automatically reconstructed 2D human poses (i.e., keypoints) and lack accurate 3D information. Furthermore, existing state-of-the-art for automatic 3D human pose estimation from sign language videos is prone to self-occlusion, noise, and motion blur effects, resulting in poor reconstruction quality. In response to this, we introduce DexAvatar, a novel framework to reconstruct bio-mechanically accurate fine-grained hand articulations and body movements from in-the-wild monocular sign language videos, guided by learned 3D hand and body priors. DexAvatar achieves strong performance in the SGNify motion capture dataset, the only benchmark available for this task, reaching an improvement of 35.11% in the estimation of body and hand poses compared to the state-of-the-art. The official website of this work is: https://github.com/kaustesseract/DexAvatar.

Method: UniPath uses Multi-Stream Control with three streams: 1) Raw-Text stream, 2) High-Level Semantics stream that uses learnable queries to a frozen pathology MLLM to extract paraphrase-robust Diagnostic Semantic Tokens and expand prompts into diagnosis-aware attribute bundles, and 3) Prototype stream for component-level morphological control via a prototype bank. The authors also curated a 2.65M image-text corpus and a 68K finely annotated subset.

Result: UniPath achieves state-of-the-art performance with Patho-FID of 80.9 (51% better than second-best) and fine-grained semantic control achieving 98.7% of real-image quality. The framework enables comprehensive four-tier evaluation tailored to pathology.

Conclusion: UniPath bridges the gap between understanding and generation in computational pathology by leveraging diagnostic semantics for controllable image generation. The framework, curated datasets, source code, and pre-trained models will be made publicly available to advance the field.

Abstract: In computational pathology, understanding and generation have evolved along disparate paths: advanced understanding models already exhibit diagnostic-level competence, whereas generative models largely simulate pixels. Progress remains hindered by three coupled factors: the scarcity of large, high-quality image-text corpora; the lack of precise, fine-grained semantic control, which forces reliance on non-semantic cues; and terminological heterogeneity, where diverse phrasings for the same diagnostic concept impede reliable text conditioning. We introduce UniPath, a semantics-driven pathology image generation framework that leverages mature diagnostic understanding to enable controllable generation. UniPath implements Multi-Stream Control: a Raw-Text stream; a High-Level Semantics stream that uses learnable queries to a frozen pathology MLLM to distill paraphrase-robust Diagnostic Semantic Tokens and to expand prompts into diagnosis-aware attribute bundles; and a Prototype stream that affords component-level morphological control via a prototype bank. On the data front, we curate a 2.65M image-text corpus and a finely annotated, high-quality 68K subset to alleviate data scarcity. For a comprehensive assessment, we establish a four-tier evaluation hierarchy tailored to pathology. Extensive experiments demonstrate UniPath’s SOTA performance, including a Patho-FID of 80.9 (51% better than the second-best) and fine-grained semantic control achieving 98.7% of the real-image. The meticulously curated datasets, complete source code, and pre-trained model weights developed in this study will be made openly accessible to the public.

Method: Proposes Decomposition and Composition framework: 1) Decomposition strategy meticulously decomposes fused multimodal features into distinct unimodal features and aligns them with ground truth unimodal counterparts; 2) Composition strategy integrates multiple unimodal features and uses them as self-supervised guidance to enhance multimodal representation learning.

Result: Extensive experiments on NTU RGB+D 60, NTU RGB+D 120, and PKU-MMD II datasets demonstrate that the proposed method achieves an excellent balance between computational cost and model performance.

Conclusion: The Decomposition and Composition framework successfully addresses the efficiency-effectiveness dilemma in multimodal action understanding by leveraging self-supervised learning to effectively utilize modality complementarity while maintaining computational efficiency.

Abstract: Multimodal human action understanding is a significant problem in computer vision, with the central challenge being the effective utilization of the complementarity among diverse modalities while maintaining model efficiency. However, most existing methods rely on simple late fusion to enhance performance, which results in substantial computational overhead. Although early fusion with a shared backbone for all modalities is efficient, it struggles to achieve excellent performance. To address the dilemma of balancing efficiency and effectiveness, we introduce a self-supervised multimodal skeleton-based action representation learning framework, named Decomposition and Composition. The Decomposition strategy meticulously decomposes the fused multimodal features into distinct unimodal features, subsequently aligning them with their respective ground truth unimodal counterparts. On the other hand, the Composition strategy integrates multiple unimodal features, leveraging them as self-supervised guidance to enhance the learning of multimodal representations. Extensive experiments on the NTU RGB+D 60, NTU RGB+D 120, and PKU-MMD II datasets demonstrate that the proposed method strikes an excellent balance between computational cost and model performance.

Method: Builds on VGGT backbone for multi-view 3D representations, adapts with feature aggregators, fine-tunes for place recognition. Uses both 3D tokens and intermediate 2D tokens with dedicated aggregation modules for 2D and 3D features. Incorporates single- and multi-frame aggregation schemes with variable-length sequence retrieval strategy.

Result: UniPR-3D sets new state-of-the-art, outperforming both single- and multi-view baselines, demonstrating effectiveness of geometry-grounded tokens for VPR.

Conclusion: The work introduces the first VPR architecture that effectively integrates multi-view information, highlighting the value of geometry-grounded tokens and showing superior generalization across diverse environments.

Abstract: Visual Place Recognition (VPR) has been traditionally formulated as a single-image retrieval task. Using multiple views offers clear advantages, yet this setting remains relatively underexplored and existing methods often struggle to generalize across diverse environments. In this work we introduce UniPR-3D, the first VPR architecture that effectively integrates information from multiple views. UniPR-3D builds on a VGGT backbone capable of encoding multi-view 3D representations, which we adapt by designing feature aggregators and fine-tune for the place recognition task. To construct our descriptor, we jointly leverage the 3D tokens and intermediate 2D tokens produced by VGGT. Based on their distinct characteristics, we design dedicated aggregation modules for 2D and 3D features, allowing our descriptor to capture fine-grained texture cues while also reasoning across viewpoints. To further enhance generalization, we incorporate both single- and multi-frame aggregation schemes, along with a variable-length sequence retrieval strategy. Our experiments show that UniPR-3D sets a new state of the art, outperforming both single- and multi-view baselines and highlighting the effectiveness of geometry-grounded tokens for VPR. Our code and models will be made publicly available on Github https://github.com/dtc111111/UniPR-3D.

Method: Multi-task model with non-causal attention to capture entire table structure, combined with parallel inference algorithm for faster cell content recognition.

Result: Demonstrated superiority both visually and statistically on two large public datasets.

Conclusion: The proposed approach effectively addresses table recognition challenges with improved efficiency and explainable performance.

Abstract: The extraction and use of diverse knowledge from numerous documents is a pressing challenge in intelligent information retrieval. Documents contain elements that require different recognition methods. Table recognition typically consists of three subtasks, namely table structure, cell position and cell content recognition. Recent models have achieved excellent recognition with a combination of multi-task learning, local attention, and mutual learning. However, their effectiveness has not been fully explained, and they require a long period of time for inference. This paper presents a novel multi-task model that utilizes non-causal attention to capture the entire table structure, and a parallel inference algorithm for faster cell content inference. The superiority is demonstrated both visually and statistically on two large public datasets.

Method: Created T2AV-Compass with 500 diverse prompts via taxonomy-driven pipeline, plus a dual-level evaluation framework integrating objective signal-level metrics (video/audio quality, cross-modal alignment) and subjective MLLM-as-a-Judge protocol.

Result: Evaluation of 11 representative T2AV systems shows even the strongest models fall substantially short of human-level realism and cross-modal consistency, with persistent failures in audio realism, fine-grained synchronization, and instruction following.

Conclusion: T2AV-Compass serves as a challenging diagnostic testbed highlighting significant improvement room for future models and advancing text-to-audio-video generation.

Abstract: Text-to-Audio-Video (T2AV) generation aims to synthesize temporally coherent video and semantically synchronized audio from natural language, yet its evaluation remains fragmented, often relying on unimodal metrics or narrowly scoped benchmarks that fail to capture cross-modal alignment, instruction following, and perceptual realism under complex prompts. To address this limitation, we present T2AV-Compass, a unified benchmark for comprehensive evaluation of T2AV systems, consisting of 500 diverse and complex prompts constructed via a taxonomy-driven pipeline to ensure semantic richness and physical plausibility. Besides, T2AV-Compass introduces a dual-level evaluation framework that integrates objective signal-level metrics for video quality, audio quality, and cross-modal alignment with a subjective MLLM-as-a-Judge protocol for instruction following and realism assessment. Extensive evaluation of 11 representative T2AVsystems reveals that even the strongest models fall substantially short of human-level realism and cross-modal consistency, with persistent failures in audio realism, fine-grained synchronization, instruction following, etc. These results indicate significant improvement room for future models and highlight the value of T2AV-Compass as a challenging and diagnostic testbed for advancing text-to-audio-video generation.

Method: Created UniRec40M dataset (40M samples), introduced hierarchical supervision training for structural comprehension, and semantic-decoupled tokenizer to separate text and formula representations.

Result: Outperforms both general-purpose VLMs and leading document parsing expert models while achieving 2-9× speedup, validated on comprehensive Chinese/English benchmarks.

Conclusion: UniRec-0.1B demonstrates that lightweight unified recognition models can achieve superior performance and efficiency for practical document parsing applications.

Abstract: Text and formulas constitute the core informational components of many documents. Accurately and efficiently recognizing both is crucial for developing robust and generalizable document parsing systems. Recently, vision-language models (VLMs) have achieved impressive unified recognition of text and formulas. However, they are large-sized and computationally demanding, restricting their usage in many applications. In this paper, we propose UniRec-0.1B, a unified recognition model with only 0.1B parameters. It is capable of performing text and formula recognition at multiple levels, including characters, words, lines, paragraphs, and documents. To implement this task, we first establish UniRec40M, a large-scale dataset comprises 40 million text, formula and their mix samples, enabling the training of a powerful yet lightweight model. Secondly, we identify two challenges when building such a lightweight but unified expert model. They are: structural variability across hierarchies and semantic entanglement between textual and formulaic content. To tackle these, we introduce a hierarchical supervision training that explicitly guides structural comprehension, and a semantic-decoupled tokenizer that separates text and formula representations. Finally, we develop a comprehensive evaluation benchmark covering Chinese and English documents from multiple domains and with multiple levels. Experimental results on this and public benchmarks demonstrate that UniRec-0.1B outperforms both general-purpose VLMs and leading document parsing expert models, while achieving a 2-9$\times$ speedup, validating its effectiveness and efficiency. Codebase and Dataset: https://github.com/Topdu/OpenOCR.

Method: FreeInpaint introduces two key techniques: 1) Prior-guided noise optimization that steers model attention toward valid inpainting regions by optimizing initial noise, and 2) A composite guidance objective tailored for inpainting that directs the denoising process by optimizing intermediate latents at each step.

Result: Extensive experiments with various inpainting diffusion models and evaluation metrics demonstrate the effectiveness and robustness of FreeInpaint in improving both prompt alignment and visual rationality.

Conclusion: FreeInpaint provides a plug-and-play, tuning-free approach that directly optimizes diffusion latents during inference to achieve better text-guided image inpainting results without requiring model retraining.

Abstract: Text-guided image inpainting endeavors to generate new content within specified regions of images using textual prompts from users. The primary challenge is to accurately align the inpainted areas with the user-provided prompts while maintaining a high degree of visual fidelity. While existing inpainting methods have produced visually convincing results by leveraging the pre-trained text-to-image diffusion models, they still struggle to uphold both prompt alignment and visual rationality simultaneously. In this work, we introduce FreeInpaint, a plug-and-play tuning-free approach that directly optimizes the diffusion latents on the fly during inference to improve the faithfulness of the generated images. Technically, we introduce a prior-guided noise optimization method that steers model attention towards valid inpainting regions by optimizing the initial noise. Furthermore, we meticulously design a composite guidance objective tailored specifically for the inpainting task. This objective efficiently directs the denoising process, enhancing prompt alignment and visual rationality by optimizing intermediate latents at each step. Through extensive experiments involving various inpainting diffusion models and evaluation metrics, we demonstrate the effectiveness and robustness of our proposed FreeInpaint.

Method: Created MarineEval, the first large-scale marine VLM benchmark with 2,000 image-based QA pairs covering 7 task dimensions and 20 capacity dimensions. Domain requirements were integrated into data construction and verified by marine experts. Evaluated 17 existing VLMs on this benchmark.

Result: Existing VLMs cannot effectively answer domain-specific marine questions, showing significant performance gaps. There’s substantial room for improvement in handling specialized domain expertise.

Conclusion: Current VLMs lack sufficient marine domain expertise despite general capabilities. MarineEval provides a benchmark to drive future research in specialized domain adaptation for VLMs.

Abstract: We have witnessed promising progress led by large language models (LLMs) and further vision language models (VLMs) in handling various queries as a general-purpose assistant. VLMs, as a bridge to connect the visual world and language corpus, receive both visual content and various text-only user instructions to generate corresponding responses. Though great success has been achieved by VLMs in various fields, in this work, we ask whether the existing VLMs can act as domain experts, accurately answering marine questions, which require significant domain expertise and address special domain challenges/requirements. To comprehensively evaluate the effectiveness and explore the boundary of existing VLMs, we construct the first large-scale marine VLM dataset and benchmark called MarineEval, with 2,000 image-based question-answering pairs. During our dataset construction, we ensure the diversity and coverage of the constructed data: 7 task dimensions and 20 capacity dimensions. The domain requirements are specially integrated into the data construction and further verified by the corresponding marine domain experts. We comprehensively benchmark 17 existing VLMs on our MarineEval and also investigate the limitations of existing models in answering marine research questions. The experimental results reveal that existing VLMs cannot effectively answer the domain-specific questions, and there is still a large room for further performance improvements. We hope our new benchmark and observations will facilitate future research. Project Page: http://marineeval.hkustvgd.com/

Method: TGC-Net introduces three components: 1) Semantic-Structural Synergy Encoder (SSE) that augments CLIP’s ViT with a CNN branch for multi-scale structural refinement, 2) Domain-Augmented Text Encoder (DATE) that injects large-language-model-derived medical knowledge, and 3) Vision-Language Calibration Module (VLCM) that refines cross-modal correspondence in a unified feature space.

Result: Experiments on five datasets across chest X-ray and thoracic CT modalities show TGC-Net achieves state-of-the-art performance with substantially fewer trainable parameters, including notable Dice gains on challenging benchmarks.

Conclusion: TGC-Net successfully addresses CLIP’s limitations for medical segmentation through parameter-efficient, task-specific adaptations, demonstrating superior performance in text-guided medical segmentation across multiple imaging modalities.

Abstract: Text-guided medical segmentation enhances segmentation accuracy by utilizing clinical reports as auxiliary information. However, existing methods typically rely on unaligned image and text encoders, which necessitate complex interaction modules for multimodal fusion. While CLIP provides a pre-aligned multimodal feature space, its direct application to medical imaging is limited by three main issues: insufficient preservation of fine-grained anatomical structures, inadequate modeling of complex clinical descriptions, and domain-specific semantic misalignment. To tackle these challenges, we propose TGC-Net, a CLIP-based framework focusing on parameter-efficient, task-specific adaptations. Specifically, it incorporates a Semantic-Structural Synergy Encoder (SSE) that augments CLIP’s ViT with a CNN branch for multi-scale structural refinement, a Domain-Augmented Text Encoder (DATE) that injects large-language-model-derived medical knowledge, and a Vision-Language Calibration Module (VLCM) that refines cross-modal correspondence in a unified feature space. Experiments on five datasets across chest X-ray and thoracic CT modalities demonstrate that TGC-Net achieves state-of-the-art performance with substantially fewer trainable parameters, including notable Dice gains on challenging benchmarks.

Method: Created ORCA benchmark with 14,647 images from 478 species, featuring 42,217 bounding box annotations and 22,321 expert-verified instance captions with fine-grained visual and textual annotations capturing morphology-oriented attributes.

Result: Evaluated 18 state-of-the-art models on three tasks (object detection, instance captioning, visual grounding), revealing key challenges including species diversity, morphological overlap, and specialized domain demands.

Conclusion: ORCA establishes a comprehensive benchmark to advance marine visual understanding research by addressing data limitations and providing systematic task formulations for the marine domain.

Abstract: Marine visual understanding is essential for monitoring and protecting marine ecosystems, enabling automatic and scalable biological surveys. However, progress is hindered by limited training data and the lack of a systematic task formulation that aligns domain-specific marine challenges with well-defined computer vision tasks, thereby limiting effective model application. To address this gap, we present ORCA, a multi-modal benchmark for marine research comprising 14,647 images from 478 species, with 42,217 bounding box annotations and 22,321 expert-verified instance captions. The dataset provides fine-grained visual and textual annotations that capture morphology-oriented attributes across diverse marine species. To catalyze methodological advances, we evaluate 18 state-of-the-art models on three tasks: object detection (closed-set and open-vocabulary), instance captioning, and visual grounding. Results highlight key challenges, including species diversity, morphological overlap, and specialized domain demands, underscoring the difficulty of marine understanding. ORCA thus establishes a comprehensive benchmark to advance research in marine domain. Project Page: http://orca.hkustvgd.com/.

Method: EFR aligns domains at two complementary levels within a self-rotated proxy feature space. It performs adaptive rotations within a parameterized Lie Group to transform both source and target features into an equivariant proxy space where alignment occurs. Learnable rotation matrices bridge domain gaps while preserving intra-domain structural information without distortion.

Result: Comprehensive experiments on various commonly used datasets demonstrate that EFR significantly enhances generative performance within the target domain compared to existing approaches.

Conclusion: EFR provides an effective adaptation strategy for few-shot image generation by addressing the fundamental distribution structure discrepancy between source and target domains through equivariant feature rotation, enabling better knowledge transfer while preserving domain structure.

Abstract: Few-shot image generation aims to effectively adapt a source generative model to a target domain using very few training images. Most existing approaches introduce consistency constraints-typically through instance-level or distribution-level loss functions-to directly align the distribution patterns of source and target domains within their respective latent spaces. However, these strategies often fall short: overly strict constraints can amplify the negative effects of the domain gap, leading to distorted or uninformative content, while overly relaxed constraints may fail to leverage the source domain effectively. This limitation primarily stems from the inherent discrepancy in the underlying distribution structures of the source and target domains. The scarcity of target samples further compounds this issue by hindering accurate estimation of the target domain’s distribution. To overcome these limitations, we propose Equivariant Feature Rotation (EFR), a novel adaptation strategy that aligns source and target domains at two complementary levels within a self-rotated proxy feature space. Specifically, we perform adaptive rotations within a parameterized Lie Group to transform both source and target features into an equivariant proxy space, where alignment is conducted. These learnable rotation matrices serve to bridge the domain gap by preserving intra-domain structural information without distortion, while the alignment optimization facilitates effective knowledge transfer from the source to the target domain. Comprehensive experiments on a variety of commonly used datasets demonstrate that our method significantly enhances the generative performance within the targeted domain.

Method: Proposes NAME framework using Implicit Neural Representations (INRs) with hierarchical temporal encoding to extract multi-scale motion features, and integrates parametric activation functions powered by Fourier transformations in MLP-based decoder for enhanced expressiveness.

Result: Extensive evaluations across several benchmark datasets demonstrate effectiveness and robustness in representing complex motion behaviors with high accuracy, enabling interpolation, inbetweening, and extrapolation at arbitrary frame rates.

Conclusion: The proposed NAME framework successfully creates continuous representations of human motion sequences with superior temporal flexibility and accuracy, advancing motion analysis and synthesis capabilities.

Abstract: Physical motions are inherently continuous, and higher camera frame rates typically contribute to improved smoothness and temporal coherence. For the first time, we explore continuous representations of human motion sequences, featuring the ability to interpolate, inbetween, and even extrapolate any input motion sequences at arbitrary frame rates. To achieve this, we propose a novel parametric activation-induced hierarchical implicit representation framework, referred to as NAME, based on Implicit Neural Representations (INRs). Our method introduces a hierarchical temporal encoding mechanism that extracts features from motion sequences at multiple temporal scales, enabling effective capture of intricate temporal patterns. Additionally, we integrate a custom parametric activation function, powered by Fourier transformations, into the MLP-based decoder to enhance the expressiveness of the continuous representation. This parametric formulation significantly augments the model’s ability to represent complex motion behaviors with high accuracy. Extensive evaluations across several benchmark datasets demonstrate the effectiveness and robustness of our proposed approach.

Method: Two-stage generation pipeline: 1) Coarse global structure synthesis, 2) Detail refinement using voxel-based refinement at fixed spatial locations with RoPE encoding. Includes comprehensive data processing with watertight processing and quality filtering.

Result: Achieves competitive performance with existing open-source methods in both data processing quality and geometry generation, despite being trained exclusively on publicly available 3D datasets with limited resources.

Conclusion: UltraShape 1.0 demonstrates effective 3D geometry generation through a scalable diffusion framework with innovative spatial localization and data processing techniques, with plans to release code and models to support future research.

Abstract: In this report, we introduce UltraShape 1.0, a scalable 3D diffusion framework for high-fidelity 3D geometry generation. The proposed approach adopts a two-stage generation pipeline: a coarse global structure is first synthesized and then refined to produce detailed, high-quality geometry. To support reliable 3D generation, we develop a comprehensive data processing pipeline that includes a novel watertight processing method and high-quality data filtering. This pipeline improves the geometric quality of publicly available 3D datasets by removing low-quality samples, filling holes, and thickening thin structures, while preserving fine-grained geometric details. To enable fine-grained geometry refinement, we decouple spatial localization from geometric detail synthesis in the diffusion process. We achieve this by performing voxel-based refinement at fixed spatial locations, where voxel queries derived from coarse geometry provide explicit positional anchors encoded via RoPE, allowing the diffusion model to focus on synthesizing local geometric details within a reduced, structured solution space. Our model is trained exclusively on publicly available 3D datasets, achieving strong geometric quality despite limited training resources. Extensive evaluations demonstrate that UltraShape 1.0 performs competitively with existing open-source methods in both data processing quality and geometry generation. All code and trained models will be released to support future research.

Method: Created VisRes Bench with 19,000+ controlled task images across three complexity levels: Level 1 tests perceptual completion and global image matching under perturbations (blur, texture, occlusion, rotation); Level 2 tests rule-based inference over single attributes (color, count, orientation); Level 3 tests compositional reasoning requiring integration of multiple visual attributes.

Result: State-of-the-art VLMs perform near random under subtle perceptual perturbations, showing clear limitations in perceptual and relational visual reasoning capacities. Models demonstrate limited abstraction beyond pattern recognition.

Conclusion: VisRes provides a unified framework for advancing abstract visual reasoning in multimodal research by isolating distinct reasoning abilities and revealing current model limitations, enabling more targeted improvements in visual reasoning capabilities.

Abstract: Vision-Language Models (VLMs) have achieved remarkable progress across tasks such as visual question answering and image captioning. Yet, the extent to which these models perform visual reasoning as opposed to relying on linguistic priors remains unclear. To address this, we introduce VisRes Bench, a benchmark designed to study visual reasoning in naturalistic settings without contextual language supervision. Analyzing model behavior across three levels of complexity, we uncover clear limitations in perceptual and relational visual reasoning capacities. VisRes isolates distinct reasoning abilities across its levels. Level 1 probes perceptual completion and global image matching under perturbations such as blur, texture changes, occlusion, and rotation; Level 2 tests rule-based inference over a single attribute (e.g., color, count, orientation); and Level 3 targets compositional reasoning that requires integrating multiple visual attributes. Across more than 19,000 controlled task images, we find that state-of-the-art VLMs perform near random under subtle perceptual perturbations, revealing limited abstraction beyond pattern recognition. We conclude by discussing how VisRes provides a unified framework for advancing abstract visual reasoning in multimodal research.

Method: Multimodal teacher-student framework integrating visual cues from egocentric cameras (forward-facing + two downward-facing) with inertial signals from consumer devices. Trained directly on real-world Ego-Elec dataset rather than synthetic data.

Result: Outperforms baseline models, demonstrating effectiveness for practical full-body motion estimation. Eliminates sim-to-real gap that constrained prior work.

Conclusion: EveryWear provides a lightweight, practical human motion capture solution using everyday wearables without calibration, enabled by real-world training data and multimodal fusion, making it suitable for daily life applications like XR interaction.

Abstract: While on-body device-based human motion estimation is crucial for applications such as XR interaction, existing methods often suffer from poor wearability, expensive hardware, and cumbersome calibration, which hinder their adoption in daily life. To address these challenges, we present EveryWear, a lightweight and practical human motion capture approach based entirely on everyday wearables: a smartphone, smartwatch, earbuds, and smart glasses equipped with one forward-facing and two downward-facing cameras, requiring no explicit calibration before use. We introduce Ego-Elec, a 9-hour real-world dataset covering 56 daily activities across 17 diverse indoor and outdoor environments, with ground-truth 3D annotations provided by the motion capture (MoCap), to facilitate robust research and benchmarking in this direction. Our approach employs a multimodal teacher-student framework that integrates visual cues from egocentric cameras with inertial signals from consumer devices. By training directly on real-world data rather than synthetic data, our model effectively eliminates the sim-to-real gap that constrains prior work. Experiments demonstrate that our method outperforms baseline models, validating its effectiveness for practical full-body motion estimation.

Method: A task-agnostic mechanism that trains LMMs to discover and use visual reasoning tokens without explicit supervision. These tokens attend globally and re-encode images in a task-adaptive way, allowing models to extract relevant visual information without hand-crafted supervision.

Result: The approach outperforms direct fine-tuning and achieves state-of-the-art results on diverse vision-centric tasks, including those where intermediate abstractions are hard to specify. It also generalizes well to multi-task instruction tuning.

Conclusion: The proposed unsupervised visual token discovery method effectively addresses the limitations of text-centric LMMs for visual reasoning, offering better performance and generalization without the need for expensive annotations or restrictive priors.

Abstract: While Large Multimodal Models (LMMs) have made significant progress, they remain largely text-centric, relying on language as their core reasoning modality. As a result, they are limited in their ability to handle reasoning tasks that are predominantly visual. Recent approaches have sought to address this by supervising intermediate visual steps with helper images, depth maps, or image crops. However, these strategies impose restrictive priors on what “useful” visual abstractions look like, add heavy annotation costs, and struggle to generalize across tasks. To address this critical limitation, we propose a task-agnostic mechanism that trains LMMs to discover and use visual reasoning tokens without explicit supervision. These tokens attend globally and re-encode the image in a task-adaptive way, enabling the model to extract relevant visual information without hand-crafted supervision. Our approach outperforms direct fine-tuning and achieves state-of-the-art results on a diverse range of vision-centric tasks – including those where intermediate abstractions are hard to specify – while also generalizing to multi-task instruction tuning.

Method: Two-stage pipeline: 1) BM25-based candidate filtering using event-centric entity extraction from captions, 2) BEiT-3-based reranking for deep multimodal semantic matching.

Result: Achieves mean average precision of 0.559 on OpenEvents v1 benchmark, substantially outperforming prior baselines.

Conclusion: Combining event-guided filtering with long-text vision-language modeling enables accurate and efficient retrieval in complex real-world scenarios.

Abstract: Retrieving images from natural language descriptions is a core task at the intersection of computer vision and natural language processing, with wide-ranging applications in search engines, media archiving, and digital content management. However, real-world image-text retrieval remains challenging due to vague or context-dependent queries, linguistic variability, and the need for scalable solutions. In this work, we propose a lightweight two-stage retrieval pipeline that leverages event-centric entity extraction to incorporate temporal and contextual signals from real-world captions. The first stage performs efficient candidate filtering using BM25 based on salient entities, while the second stage applies BEiT-3 models to capture deep multimodal semantics and rerank the results. Evaluated on the OpenEvents v1 benchmark, our method achieves a mean average precision of 0.559, substantially outperforming prior baselines. These results highlight the effectiveness of combining event-guided filtering with long-text vision-language modeling for accurate and efficient retrieval in complex, real-world scenarios. Our code is available at https://github.com/PhamPhuHoa-23/Event-Based-Image-Retrieval

Method: Three modules: (1) Text Segment Extraction - decomposes complex descriptions into temporally ordered atomic action phrases; (2) Motion Segment Extraction - partitions motion sequences into corresponding segments; (3) Fine-grained Text-Motion Alignment - aligns text and motion segments using contrastive learning.

Result: Improves strong baseline on two widely used datasets, achieving TOP 1 score of 0.553 on HumanML3D test set. Learned shared embedding space enables application to retrieval tasks like motion grounding and motion-to-text retrieval.

Conclusion: Segment-level alignment enables finer-grained text-motion correspondence, improving generation quality and enabling cross-modal retrieval applications through learned shared embedding space.

Abstract: Generating 3D human motions from textual descriptions is an important research problem with broad applications in video games, virtual reality, and augmented reality. Recent methods align the textual description with human motion at the sequence level, neglecting the internal semantic structure of modalities. However, both motion descriptions and motion sequences can be naturally decomposed into smaller and semantically coherent segments, which can serve as atomic alignment units to achieve finer-grained correspondence. Motivated by this, we propose SegMo, a novel Segment-aligned text-conditioned human Motion generation framework to achieve fine-grained text-motion alignment. Our framework consists of three modules: (1) Text Segment Extraction, which decomposes complex textual descriptions into temporally ordered phrases, each representing a simple atomic action; (2) Motion Segment Extraction, which partitions complete motion sequences into corresponding motion segments; and (3) Fine-grained Text-Motion Alignment, which aligns text and motion segments with contrastive learning. Extensive experiments demonstrate that SegMo improves the strong baseline on two widely used datasets, achieving an improved TOP 1 score of 0.553 on the HumanML3D test set. Moreover, thanks to the learned shared embedding space for text and motion segments, SegMo can also be applied to retrieval-style tasks such as motion grounding and motion-to-text retrieval.

Method: Proposes a two-stage modular architecture: (1) Causal Chain Extractor (CCE) generates natural language causal chains from video-question pairs, and (2) Causal Chain-Driven Answerer (CCDA) derives answers grounded in these chains. Introduces scalable method to generate annotated causal chains from existing datasets and creates human-verified chains for 46K samples.

Result: Outperforms state-of-the-art models on three large-scale benchmarks, yields substantial gains in explainability, user trust, and generalization. CCE serves as reusable causal reasoning engine across domains.

Conclusion: The modular approach with explicit causal chain representations enables transparent, logically coherent inference, addressing limitations of black-box VideoQA models while improving both performance and interpretability.

Abstract: Existing Causal-Why Video Question Answering (VideoQA) models often struggle with higher-order reasoning, relying on opaque, monolithic pipelines that entangle video understanding, causal inference, and answer generation. These black-box approaches offer limited interpretability and tend to depend on shallow heuristics. We propose a novel, modular paradigm that explicitly decouples causal reasoning from answer generation, introducing natural language causal chains as interpretable intermediate representations. Inspired by human cognitive models, these structured cause-effect sequences bridge low-level video content with high-level causal reasoning, enabling transparent and logically coherent inference. Our two-stage architecture comprises a Causal Chain Extractor (CCE) that generates causal chains from video-question pairs, and a Causal Chain-Driven Answerer (CCDA) that derives answers grounded in these chains. To address the lack of annotated reasoning traces, we introduce a scalable method for generating accurate causal chains from existing datasets. We construct human verified causal chains for 46K samples. We also propose CauCo, a new evaluation metric for causality-oriented captioning. Experiments on three large-scale benchmarks demonstrate that our approach not only outperforms state-of-the-art models, but also yields substantial gains in explainability, user trust, and generalization – positioning the CCE as a reusable causal reasoning engine across diverse domains. Project page: https://paritoshparmar.github.io/chainreaction/

Method: Three-pronged approach: 1) Lightweight intermediate-conditioning mechanism in DiT architecture with Adaptive Tuning for arbitrary-frame control; 2) High-quality dataset curation with Visual Expression SFT and Tailored DPO for motion rationality and transition smoothness; 3) Segment-wise Auto-Regressive (SAR) inference for memory-efficient long sequence generation.

Result: Extensive experiments show the approach achieves visually striking and seamlessly coherent one-shot effects while maintaining computational efficiency, enabling transformation of fragmented visual materials into cohesive cinematic experiences.

Conclusion: DreaMontage provides a comprehensive framework for generating high-quality, expressive one-shot videos from diverse user inputs, overcoming limitations of existing methods and making one-shot filmmaking more accessible through virtual generation.

Abstract: The “one-shot” technique represents a distinct and sophisticated aesthetic in filmmaking. However, its practical realization is often hindered by prohibitive costs and complex real-world constraints. Although emerging video generation models offer a virtual alternative, existing approaches typically rely on naive clip concatenation, which frequently fails to maintain visual smoothness and temporal coherence. In this paper, we introduce DreaMontage, a comprehensive framework designed for arbitrary frame-guided generation, capable of synthesizing seamless, expressive, and long-duration one-shot videos from diverse user-provided inputs. To achieve this, we address the challenge through three primary dimensions. (i) We integrate a lightweight intermediate-conditioning mechanism into the DiT architecture. By employing an Adaptive Tuning strategy that effectively leverages base training data, we unlock robust arbitrary-frame control capabilities. (ii) To enhance visual fidelity and cinematic expressiveness, we curate a high-quality dataset and implement a Visual Expression SFT stage. In addressing critical issues such as subject motion rationality and transition smoothness, we apply a Tailored DPO scheme, which significantly improves the success rate and usability of the generated content. (iii) To facilitate the production of extended sequences, we design a Segment-wise Auto-Regressive (SAR) inference strategy that operates in a memory-efficient manner. Extensive experiments demonstrate that our approach achieves visually striking and seamlessly coherent one-shot effects while maintaining computational efficiency, empowering users to transform fragmented visual materials into vivid, cohesive one-shot cinematic experiences.

Method: Created large-scale RSCC dataset with 62,351 pre-/post-disaster image pairs covering earthquakes, floods, wildfires, etc., paired with human-like change captions.

Result: RSCC enables robust training/evaluation of vision-language models for disaster-aware bi-temporal understanding and facilitates detailed disaster-related analysis.

Conclusion: RSCC bridges temporal and semantic gaps in remote sensing data, paving way for more accurate, interpretable, and scalable vision-language applications.

Abstract: Remote sensing is critical for disaster monitoring, yet existing datasets lack temporal image pairs and detailed textual annotations. While single-snapshot imagery dominates current resources, it fails to capture dynamic disaster impacts over time. To address this gap, we introduce the Remote Sensing Change Caption (RSCC) dataset, a large-scale benchmark comprising 62,351 pre-/post-disaster image pairs (spanning earthquakes, floods, wildfires, and more) paired with rich, human-like change captions. By bridging the temporal and semantic divide in remote sensing data, RSCC enables robust training and evaluation of vision-language models for disaster-aware bi-temporal understanding. Our results highlight RSCC’s ability to facilitate detailed disaster-related analysis, paving the way for more accurate, interpretable, and scalable vision-language applications in remote sensing. Code and dataset are available at https://github.com/Bili-Sakura/RSCC.

Method: 1) Dual-pathway DINOv2 encoder with feature distribution alignment mechanism that statistically aligns incomplete-modality features with full-modality representations; 2) Intrinsic Normal Prototypes (INPs) extractor and INP-guided decoder that reconstruct only normal anatomical patterns while amplifying abnormal deviations; 3) Randomized modality masking and indirect feature completion during training to adapt to all modality configurations without re-training.

Result: Extensive experiments on BraTS2018, MU-Glioma-Post, and Pretreat-MetsToBrain-Masks show the approach consistently surpasses state-of-the-art industrial and medical AD baselines across 7 modality combinations, achieving superior generalization.

Conclusion: The study establishes a scalable paradigm for multimodal medical anomaly detection under real-world, imperfect modality conditions, enabling robust performance with arbitrary MRI modality availability.

Abstract: Reliable anomaly detection in brain MRI remains challenging due to the scarcity of annotated abnormal cases and the frequent absence of key imaging modalities in real clinical workflows. Existing single-class or multi-class anomaly detection (AD) models typically rely on fixed modality configurations, require repetitive training, or fail to generalize to unseen modality combinations, limiting their clinical scalability. In this work, we present a unified Any-Modality AD framework that performs robust anomaly detection and localization under arbitrary MRI modality availability. The framework integrates a dual-pathway DINOv2 encoder with a feature distribution alignment mechanism that statistically aligns incomplete-modality features with full-modality representations, enabling stable inference even with severe modality dropout. To further enhance semantic consistency, we introduce an Intrinsic Normal Prototypes (INPs) extractor and an INP-guided decoder that reconstruct only normal anatomical patterns while naturally amplifying abnormal deviations. Through randomized modality masking and indirect feature completion during training, the model learns to adapt to all modality configurations without re-training. Extensive experiments on BraTS2018, MU-Glioma-Post, and Pretreat-MetsToBrain-Masks demonstrate that our approach consistently surpasses state-of-the-art industrial and medical AD baselines across 7 modality combinations, achieving superior generalization. This study establishes a scalable paradigm for multimodal medical AD under real-world, imperfect modality conditions. Our source code is available at https://github.com/wuchangw/AnyAD.

Method: Develop scalable 2DGS pipeline with structured initialization, luminance-aware pruning, and batched CUDA kernels. Adapt CLIP training to 2DGS by reusing frozen RGB transformer backbone with lightweight splat-aware input stem and perceiver resampler, training only 9.7-13.8% of parameters.

Result: Achieved 90x faster fitting and ~97% GPU utilization compared to prior implementations. GS encoders yield competitive zero-shot performance on 38 CLIP benchmark datasets while compressing inputs 3x to 23.5x relative to pixels.

Conclusion: 2DGS is established as a viable multimodal substrate that addresses architectural bottlenecks and opens a path toward representations that are both semantically powerful and transmission-efficient for edge-cloud learning.

Abstract: Modern vision language pipelines are driven by RGB vision encoders trained on massive image text corpora. While these pipelines have enabled impressive zero-shot capabilities and strong transfer across tasks, they still inherit two structural inefficiencies from the pixel domain: (i) transmitting dense RGB images from edge devices to the cloud is energy-intensive and costly, and (ii) patch-based tokenization explodes sequence length, stressing attention budgets and context limits. We explore 2D Gaussian Splatting (2DGS) as an alternative visual substrate for alignment: a compact, spatially adaptive representation that parameterizes images by a set of colored anisotropic Gaussians. We develop a scalable 2DGS pipeline with structured initialization, luminance-aware pruning, and batched CUDA kernels, achieving over 90x faster fitting and about 97% GPU utilization compared to prior implementations. We further adapt contrastive language-image pre-training (CLIP) to 2DGS by reusing a frozen RGB-based transformer backbone with a lightweight splat-aware input stem and a perceiver resampler, training only 9.7% to 13.8% of the total parameters. On a 12.8M dataset from DataComp, GS encoders yield competitive zero-shot performance on 38 datasets from the CLIP benchmark while compressing inputs 3x to 23.5x relative to pixels. Our results establish 2DGS as a viable multimodal substrate, pinpoint architectural bottlenecks, and open a path toward representations that are both semantically powerful and transmission-efficient for edge-cloud learning.

Method: Proposes Attention-Conditional Diffusion (ACD) that aligns model attention maps with external control signals. Introduces sparse 3D-aware object layout as conditioning signal, Layout ControlNet for processing, and automated annotation pipeline for scalable layout integration.

Result: Extensive experiments show ACD delivers superior alignment with conditioning inputs while preserving temporal coherence and visual fidelity, establishing an effective paradigm for conditional video synthesis.

Conclusion: ACD provides a novel framework for direct conditional control in video diffusion models through attention supervision, overcoming limitations of existing guidance methods and achieving better controllability.

Abstract: Controllability is a fundamental requirement in video synthesis, where accurate alignment with conditioning signals is essential. Existing classifier-free guidance methods typically achieve conditioning indirectly by modeling the joint distribution of data and conditions, which often results in limited controllability over the specified conditions. Classifier-based guidance enforces conditions through an external classifier, but the model may exploit this mechanism to raise the classifier score without genuinely satisfying the intended condition, resulting in adversarial artifacts and limited effective controllability. In this paper, we propose Attention-Conditional Diffusion (ACD), a novel framework for direct conditional control in video diffusion models via attention supervision. By aligning the model’s attention maps with external control signals, ACD achieves better controllability. To support this, we introduce a sparse 3D-aware object layout as an efficient conditioning signal, along with a dedicated Layout ControlNet and an automated annotation pipeline for scalable layout integration. Extensive experiments on benchmark video generation datasets demonstrate that ACD delivers superior alignment with conditioning inputs while preserving temporal coherence and visual fidelity, establishing an effective paradigm for conditional video synthesis.

Method: Factorizes generation into two stages: 1) Train a generative model on grid images of subsampled frames to generate coarse sequences at low resolution using Diffusion Transformer (DiT) to capture frame correlations, effectively extending 2D image generation to low-resolution 3D sequences without architectural changes. 2) Super-resolve each frame individually to add high-resolution details.

Result: Achieves superior synthesis quality and improved coherence across sequences compared to existing methods. Enables high-fidelity generation of arbitrary-length sequences with increased efficiency in inference time (at least twice-as-fast) and training data usage. Generalizes effectively across diverse data domains without requiring additional priors or supervision.

Conclusion: The proposed factorization approach overcomes key limitations of SoTA image sequence generation methods, offering a more effective representation that consistently outperforms existing methods in both quality and inference speed across diverse datasets.

Abstract: Modern deep learning methods typically treat image sequences as large tensors of sequentially stacked frames. However, is this straightforward representation ideal given the current state-of-the-art (SoTA)? In this work, we address this question in the context of generative models and aim to devise a more effective way of modeling image sequence data. Observing the inefficiencies and bottlenecks of current SoTA image sequence generation methods, we showcase that rather than working with large tensors, we can improve the generation process by factorizing it into first generating the coarse sequence at low resolution and then refining the individual frames at high resolution. We train a generative model solely on grid images comprising subsampled frames. Yet, we learn to generate image sequences, using the strong self-attention mechanism of the Diffusion Transformer (DiT) to capture correlations between frames. In effect, our formulation extends a 2D image generator to operate as a low-resolution 3D image-sequence generator without introducing any architectural modifications. Subsequently, we super-resolve each frame individually to add the sequence-independent high-resolution details. This approach offers several advantages and can overcome key limitations of the SoTA in this domain. Compared to existing image sequence generation models, our method achieves superior synthesis quality and improved coherence across sequences. It also delivers high-fidelity generation of arbitrary-length sequences and increased efficiency in inference time and training data usage. Furthermore, our straightforward formulation enables our method to generalize effectively across diverse data domains, which typically require additional priors and supervision to model in a generative context. Our method consistently outperforms SoTA in quality and inference speed (at least twice-as-fast) across datasets.

Method: Proposes SpikeSurgSeg framework with: 1) Surgical-scene masked autoencoding pretraining for SNNs using layer-wise tube masking for robust spatiotemporal representation learning, and 2) Lightweight spike-driven segmentation head for temporally consistent predictions while maintaining low latency.

Result: Achieves mIoU comparable to SOTA ANN-based models while reducing inference latency by at least 8x and delivering over 20x acceleration relative to foundation-model baselines on EndoVis18 and SurgBleed datasets.

Conclusion: SpikeSurgSeg demonstrates the potential of SNNs for time-critical surgical scene segmentation, offering efficient real-time performance on non-GPU platforms while maintaining competitive accuracy with conventional deep learning models.

Abstract: Modern surgical systems increasingly rely on intelligent scene understanding to provide timely situational awareness for enhanced intra-operative safety. Within this pipeline, surgical scene segmentation plays a central role in accurately perceiving operative events. Although recent deep learning models, particularly large-scale foundation models, achieve remarkable segmentation accuracy, their substantial computational demands and power consumption hinder real-time deployment in resource-constrained surgical environments. To address this limitation, we explore the emerging SNN as a promising paradigm for highly efficient surgical intelligence. However, their performance is still constrained by the scarcity of labeled surgical data and the inherently sparse nature of surgical video representations. To this end, we propose \textit{SpikeSurgSeg}, the first spike-driven video Transformer framework tailored for surgical scene segmentation with real-time potential on non-GPU platforms. To address the limited availability of surgical annotations, we introduce a surgical-scene masked autoencoding pretraining strategy for SNNs that enables robust spatiotemporal representation learning via layer-wise tube masking. Building on this pretrained backbone, we further adopt a lightweight spike-driven segmentation head that produces temporally consistent predictions while preserving the low-latency characteristics of SNNs. Extensive experiments on EndoVis18 and our in-house SurgBleed dataset demonstrate that SpikeSurgSeg achieves mIoU comparable to SOTA ANN-based models while reducing inference latency by at least $8\times$. Notably, it delivers over $20\times$ acceleration relative to most foundation-model baselines, underscoring its potential for time-critical surgical scene segmentation.

Method: Two key contributions: (1) a new large-scale dataset of image-sketch-instruction triplets for pretraining and instruction tuning, and (2) O3SLM, an LVLM trained on this dataset. The model is evaluated on multiple sketch-based tasks using existing sketch datasets (QuickDraw!, Sketchy, Tu Berlin) plus their generated SketchVCL dataset.

Result: O3SLM achieves state-of-the-art performance on sketch-based tasks including object localization, counting, image retrieval (SBIR and fine-grained SBIR), and visual question answering (VQA), substantially outperforming existing LVLMs in sketch comprehension and reasoning.

Conclusion: The proposed dataset and O3SLM model effectively address the limitations of current LVLMs in understanding abstract visual inputs like sketches, demonstrating superior performance across multiple sketch-based reasoning tasks.

Abstract: While Large Vision Language Models (LVLMs) are increasingly deployed in real-world applications, their ability to interpret abstract visual inputs remains limited. Specifically, they struggle to comprehend hand-drawn sketches, a modality that offers an intuitive means of expressing concepts that are difficult to describe textually. We identify the primary bottleneck as the absence of a large-scale dataset that jointly models sketches, photorealistic images, and corresponding natural language instructions. To address this, we present two key contributions: (1) a new, large-scale dataset of image-sketch-instruction triplets designed to facilitate both pretraining and instruction tuning, and (2) O3SLM, an LVLM trained on this dataset. Comprehensive evaluations on multiple sketch-based tasks: (a) object localization, (b) counting, (c) image retrieval i.e., (SBIR and fine-grained SBIR), and (d) visual question answering (VQA); while incorporating the three existing sketch datasets, namely QuickDraw!, Sketchy, and Tu Berlin, along with our generated SketchVCL dataset, show that O3SLM achieves state-of-the-art performance, substantially outperforming existing LVLMs in sketch comprehension and reasoning.

Method: Proposes modeling row/column transitions as 1D signals, processing them with Gaussian convolution with progressively increasing variances, followed by statistical thresholding to suppress noise while preserving structural edges. Detected peaks are mapped back to image coordinates to obtain accurate segment boundaries.

Result: Improves Cell-Aware Segmentation Accuracy (CASA) from 67% to 76% on PubLayNet-1M benchmark when using TableNet with PyTesseract OCR. Method is robust to resolution variations through zero-padding and scaling strategies.

Conclusion: The proposed multi-scale signal-processing approach provides robust table edge detection from masks, producing optimized structured tabular outputs suitable for downstream analysis in challenging document image conditions.

Abstract: Structured data extraction from tables plays a crucial role in document image analysis for scanned documents and digital archives. Although many methods have been proposed to detect table structures and extract cell contents, accurately identifying table segment boundaries (rows and columns) remains challenging, particularly in low-resolution or noisy images. In many real-world scenarios, table data are incomplete or degraded, limiting the adaptability of transformer-based methods to noisy inputs. Mask-based edge detection techniques have shown greater robustness under such conditions, as their sensitivity can be adjusted through threshold tuning; however, existing approaches typically apply masks directly to images, leading to noise sensitivity, resolution loss, or high computational cost. This paper proposes a novel multi-scale signal-processing method for detecting table edges from table masks. Row and column transitions are modeled as one-dimensional signals and processed using Gaussian convolution with progressively increasing variances, followed by statistical thresholding to suppress noise while preserving stable structural edges. Detected signal peaks are mapped back to image coordinates to obtain accurate segment boundaries. Experimental results show that applying the proposed approach to column edge detection improves Cell-Aware Segmentation Accuracy (CASA) a layout-aware metric evaluating both textual correctness and correct cell placement from 67% to 76% on the PubLayNet-1M benchmark when using TableNet with PyTesseract OCR. The method is robust to resolution variations through zero-padding and scaling strategies and produces optimized structured tabular outputs suitable for downstream analysis.

Method: AndroidLens framework includes: (1) 571 long-latency tasks (avg >26 steps) from real-world scenarios across 38 domains, (2) static evaluation preserving real-world anomalies with multiple valid paths, and (3) dynamic evaluation using milestone-based scheme with Average Task Progress (ATP) metric for fine-grained measurement.

Result: Even the best models achieve only 12.7% task success rate and 50.47% ATP, highlighting the difficulty of real-world GUI automation. Key challenges identified include environmental anomalies, adaptive exploration, and long-term memory retention.

Conclusion: AndroidLens provides a comprehensive, challenging benchmark for mobile GUI agents that better reflects real-world complexity. The low performance of current models reveals significant gaps in handling realistic mobile automation tasks, pointing to important research directions for improving GUI agents.

Abstract: Graphical user interface (GUI) agents can substantially improve productivity by automating frequently executed long-latency tasks on mobile devices. However, existing evaluation benchmarks are still constrained to limited applications, simple tasks, and coarse-grained metrics. To address this, we introduce AndroidLens, a challenging evaluation framework for mobile GUI agents, comprising 571 long-latency tasks in both Chinese and English environments, each requiring an average of more than 26 steps to complete. The framework features: (1) tasks derived from real-world user scenarios across 38 domains, covering complex types such as multi-constraint, multi-goal, and domain-specific tasks; (2) static evaluation that preserves real-world anomalies and allows multiple valid paths to reduce bias; and (3) dynamic evaluation that employs a milestone-based scheme for fine-grained progress measurement via Average Task Progress (ATP). Our evaluation indicates that even the best models reach only a 12.7% task success rate and 50.47% ATP. We also underscore key challenges in real-world environments, including environmental anomalies, adaptive exploration, and long-term memory retention.

Method: TICON uses a single shared transformer encoder pretrained with masked modeling objective to unify and contextualize representations from diverse tile-level pathology foundation models.

Result: TICON-contextualized embeddings significantly improve performance across tasks, achieving SOTA on tile-level benchmarks (HEST-Bench, THUNDER, CATCH) and slide-level benchmarks (Patho-Bench). A slide-level foundation model pretrained with only 11K WSIs outperforms SOTA models trained with up to 350K WSIs.

Conclusion: TICON provides an effective approach to contextualize tile embeddings from any foundation model, enabling better pathology analysis with less data and establishing new state-of-the-art performance across multiple benchmarks.

Abstract: The interpretation of small tiles in large whole slide images (WSI) often needs a larger image context. We introduce TICON, a transformer-based tile representation contextualizer that produces rich, contextualized embeddings for ‘‘any’’ application in computational pathology. Standard tile encoder-based pipelines, which extract embeddings of tiles stripped from their context, fail to model the rich slide-level information essential for both local and global tasks. Furthermore, different tile-encoders excel at different downstream tasks. Therefore, a unified model is needed to contextualize embeddings derived from ‘‘any’’ tile-level foundation model. TICON addresses this need with a single, shared encoder, pretrained using a masked modeling objective to simultaneously unify and contextualize representations from diverse tile-level pathology foundation models. Our experiments demonstrate that TICON-contextualized embeddings significantly improve performance across many different tasks, establishing new state-of-the-art results on tile-level benchmarks (i.e., HEST-Bench, THUNDER, CATCH) and slide-level benchmarks (i.e., Patho-Bench). Finally, we pretrain an aggregator on TICON to form a slide-level foundation model, using only 11K WSIs, outperforming SoTA slide-level foundation models pretrained with up to 350K WSIs.

Method: A text-guided token pruning framework that operates after visual encoding and before memory-based propagation. It ranks tokens using a lightweight routing mechanism integrating local visual context, semantic relevance from object-centric textual descriptions (user-provided or auto-generated), and uncertainty cues to preserve ambiguous/boundary regions.

Result: Achieves up to 42.50% faster inference and 37.41% lower GPU memory usage compared to unpruned baseline SAM2, while preserving competitive J and F performance across multiple challenging video segmentation benchmarks.

Conclusion: Post-encoder token pruning provides a practical and effective pathway to efficient, prompt-aware video segmentation, highlighting the potential of early token selection to improve transformer-based video segmentation scalability for real-time and resource-constrained applications.

Abstract: Segment Anything Model 2 (SAM2), a vision foundation model has significantly advanced in prompt-driven video object segmentation, yet their practical deployment remains limited by the high computational and memory cost of processing dense visual tokens across time. The SAM2 pipelines typically propagate all visual tokens produced by the image encoder through downstream temporal reasoning modules, regardless of their relevance to the target object, resulting in reduced scalability due to quadratic memory attention overhead. In this work, we introduce a text-guided token pruning framework that improves inference efficiency by selectively reducing token density prior to temporal propagation, without modifying the underlying segmentation architecture. Operating after visual encoding and before memory based propagation, our method ranks tokens using a lightweight routing mechanism that integrates local visual context, semantic relevance derived from object-centric textual descriptions (either user-provided or automatically generated), and uncertainty cues that help preserve ambiguous or boundary critical regions. By retaining only the most informative tokens for downstream processing, the proposed approach reduces redundant computation while maintaining segmentation fidelity. Extensive experiments across multiple challenging video segmentation benchmarks demonstrate that post-encoder token pruning provides a practical and effective pathway to efficient, prompt-aware video segmentation, achieving up to 42.50 percent faster inference and 37.41 percent lower GPU memory usage compared to the unpruned baseline SAM2, while preserving competitive J and F performance. These results highlight the potential of early token selection to improve the scalability of transformer-based video segmentation systems for real-time and resource-constrained applications.

Method: Created Streamo-Instruct-465K, a large-scale instruction-following dataset for streaming video understanding with diverse temporal contexts and multi-task supervision. Trained end-to-end through a streamlined pipeline to enable unified training across heterogeneous streaming tasks.

Result: Streamo exhibits strong temporal reasoning, responsive interaction, and broad generalization across various streaming benchmarks. It bridges the gap between offline video perception models and real-time multimodal assistants.

Conclusion: Streamo represents a step toward unified, intelligent video understanding in continuous video streams, demonstrating the potential for general-purpose interactive video assistants.

Abstract: We present Streamo, a real-time streaming video LLM that serves as a general-purpose interactive assistant. Unlike existing online video models that focus narrowly on question answering or captioning, Streamo performs a broad spectrum of streaming video tasks, including real-time narration, action understanding, event captioning, temporal event grounding, and time-sensitive question answering. To develop such versatility, we construct Streamo-Instruct-465K, a large-scale instruction-following dataset tailored for streaming video understanding. The dataset covers diverse temporal contexts and multi-task supervision, enabling unified training across heterogeneous streaming tasks. After training end-to-end on the instruction-following dataset through a streamlined pipeline, Streamo exhibits strong temporal reasoning, responsive interaction, and broad generalization across a variety of streaming benchmarks. Extensive experiments show that Streamo bridges the gap between offline video perception models and real-time multimodal assistants, making a step toward unified, intelligent video understanding in continuous video streams.

Method: Created YearGuessr dataset (55,546 building images with multi-modal attributes), framed construction year prediction as ordinal regression, introduced popularity-aware interval accuracy metrics, and benchmarked 30+ models including YearCLIP.

Result: VLMs achieve up to 34% higher accuracy on famous buildings, confirming they excel on popular/memorized items but struggle significantly with unrecognized subjects, exposing critical flaws in reasoning capabilities.

Conclusion: Current VLMs have significant popularity bias and rely on memorization rather than generalizable understanding, highlighting a critical limitation in their reasoning capabilities that needs to be addressed.

Abstract: We expose a significant popularity bias in state-of-the-art vision-language models (VLMs), which achieve up to 34% higher accuracy on famous buildings compared to ordinary ones, indicating a reliance on memorization over generalizable understanding. To systematically investigate this, we introduce the largest open benchmark for this task: the YearGuessr dataset, a collection of 55,546 building images with multi-modal attributes from 157 countries, annotated with continuous ordinal labels of their construction year (1001-2024), GPS data, and page-view counts as a proxy for popularity. Using this dataset, we frame the construction year prediction task as ordinal regression and introduce popularity-aware interval accuracy metrics to quantify this bias. Our resulting benchmark of 30+ models, including our YearCLIP model, confirms that VLMs excel on popular, memorized items but struggle significantly with unrecognized subjects, exposing a critical flaw in their reasoning capabilities. Project page: https://sytwu.github.io/BeyondMemo/

Method: HiStream uses three compression strategies: 1) Spatial Compression - denoising at low resolution first then refining at high resolution with cached features; 2) Temporal Compression - chunk-by-chunk processing with fixed-size anchor cache for stable inference; 3) Timestep Compression - applying fewer denoising steps to subsequent cache-conditioned chunks.

Result: On 1080p benchmarks, HiStream achieves state-of-the-art visual quality with 76.2x faster denoising than Wan2.1 baseline and negligible quality loss. HiStream+ (with all three optimizations) achieves 107.5x acceleration, offering compelling speed-quality trade-off.

Conclusion: HiStream makes high-resolution video generation practical and scalable by dramatically reducing computational complexity while maintaining visual quality, enabling efficient video generation for real-world applications.

Abstract: High-resolution video generation, while crucial for digital media and film, is computationally bottlenecked by the quadratic complexity of diffusion models, making practical inference infeasible. To address this, we introduce HiStream, an efficient autoregressive framework that systematically reduces redundancy across three axes: i) Spatial Compression: denoising at low resolution before refining at high resolution with cached features; ii) Temporal Compression: a chunk-by-chunk strategy with a fixed-size anchor cache, ensuring stable inference speed; and iii) Timestep Compression: applying fewer denoising steps to subsequent, cache-conditioned chunks. On 1080p benchmarks, our primary HiStream model (i+ii) achieves state-of-the-art visual quality while demonstrating up to 76.2x faster denoising compared to the Wan2.1 baseline and negligible quality loss. Our faster variant, HiStream+, applies all three optimizations (i+ii+iii), achieving a 107.5x acceleration over the baseline, offering a compelling trade-off between speed and quality, thereby making high-resolution video generation both practical and scalable.

Method: DenseVLM learns unbiased region-language alignment by leveraging pre-trained VLM to retrieve categories for unlabeled regions and decoupling interference between foreground and background features.

Result: DenseVLM can directly replace original VLMs in open-vocabulary object detection and image segmentation methods, leading to notable performance improvements, and shows promising zero-shot scalability on more extensive datasets.

Conclusion: DenseVLM effectively addresses foreground bias in VLMs for dense prediction tasks, improving performance in open-vocabulary detection and segmentation while maintaining scalability.

Abstract: Pre-trained vision-language models (VLMs), such as CLIP, have demonstrated impressive zero-shot recognition capability, but still underperform in dense prediction tasks. Self-distillation recently is emerging as a promising approach for fine-tuning VLMs to better adapt to local regions without requiring extensive annotations. However, previous state-of-the-art approaches often suffer from significant `foreground bias’, where models tend to wrongly identify background regions as foreground objects. To alleviate this issue, we propose DenseVLM, a framework designed to learn unbiased region-language alignment from powerful pre-trained VLM representations. To alleviate this issue, we propose DenseVLM, a framework designed to learn unbiased region-language alignment from powerful pre-trained VLM representations. DenseVLM leverages the pre-trained VLM to retrieve categories for unlabeled regions and then decouples the interference between foreground and background features. We show that DenseVLM can directly replace the original VLM in open-vocabulary object detection and image segmentation methods, leading to notable performance improvements. Furthermore, it exhibits promising zero-shot scalability when training on more extensive and diverse datasets. Our code is available at https://github.com/HVision-NKU/DenseVLM.

Method: Uses point maps (rasterized 3D scene coordinates) as conditioning signal for pre-trained 2D diffusion models. Incorporates reference attention layers and ControlNet for point map features to guide image generation while maintaining geometric fidelity.

Result: Achieves high-quality novel view synthesis on real-life driving data. Works flexibly with various point map conditioning signals (dense depth maps or sparse LiDAR points). Can distill to 3D representations like 3D Gaussian Splatting for improved view extrapolation.

Conclusion: PointmapDiff effectively addresses novel view synthesis in urban driving scenes by leveraging geometric priors through point maps and diffusion models, offering flexibility and high-quality results for view extrapolation tasks.

Abstract: Synthesizing extrapolated views remains a difficult task, especially in urban driving scenes, where the only reliable sources of data are limited RGB captures and sparse LiDAR points. To address this problem, we present PointmapDiff, a framework for novel view synthesis that utilizes pre-trained 2D diffusion models. Our method leverages point maps (i.e., rasterized 3D scene coordinates) as a conditioning signal, capturing geometric and photometric priors from the reference images to guide the image generation process. With the proposed reference attention layers and ControlNet for point map features, PointmapDiff can generate accurate and consistent results across varying viewpoints while respecting geometric fidelity. Experiments on real-life driving data demonstrate that our method achieves high-quality generation with flexibility over point map conditioning signals (e.g., dense depth map or even sparse LiDAR points) and can be used to distill to 3D representations such as 3D Gaussian Splatting for improving view extrapolation.

Method: Represent each ground image pixel as a 3D Gaussian with semantic/spatial features, synthesize into BEV feature map for pose estimation, and use icosphere-based supervision for panoramic queries.

Result: Significantly improves localization accuracy over prior approaches on KITTI and VIGOR datasets with both pinhole and panoramic query images.

Conclusion: BevSplat effectively resolves height ambiguity in weakly supervised cross-view localization through feature-based Gaussian primitives, outperforming existing methods.

Abstract: This paper addresses the problem of weakly supervised cross-view localization, where the goal is to estimate the pose of a ground camera relative to a satellite image with noisy ground truth annotations. A common approach to bridge the cross-view domain gap for pose estimation is Bird’s-Eye View (BEV) synthesis. However, existing methods struggle with height ambiguity due to the lack of depth information in ground images and satellite height maps. Previous solutions either assume a flat ground plane or rely on complex models, such as cross-view transformers. We propose BevSplat, a novel method that resolves height ambiguity by using feature-based Gaussian primitives. Each pixel in the ground image is represented by a 3D Gaussian with semantic and spatial features, which are synthesized into a BEV feature map for relative pose estimation. Additionally, to address challenges with panoramic query images, we introduce an icosphere-based supervision strategy for the Gaussian primitives. We validate our method on the widely used KITTI and VIGOR datasets, which include both pinhole and panoramic query images. Experimental results show that BevSplat significantly improves localization accuracy over prior approaches.

Method: Proposes a VLM-based pseudo-labeling approach with state-change dynamics constraints, and introduces the WhereToChange benchmark built on in-the-wild Internet videos.

Result: SOTA VLMs and video segmentation methods struggle with this task, validating its difficulty. The proposed model shows promise for localizing where and how fast objects change in video.

Conclusion: Spatial OSC segmentation is a new frontier task that challenges current methods and invites the community to build more state-change-sensitive representations, with applications for robotic activity progress tracking.

Abstract: Object state changes in video reveal critical cues about human and agent activity. However, existing methods are limited to temporal localization of when the object is in its initial state (e.g., cheese block) versus when it has completed a state change (e.g., grated cheese), offering no insight into where the change is unfolding. We propose to deepen the problem by introducing the spatially-progressing object state change segmentation task. The goal is to segment at the pixel-level those regions of an object that are actionable and those that are transformed. We show that state-of-the-art VLMs and video segmentation methods struggle at this task, underscoring its difficulty and novelty. As an initial baseline, we design a VLM-based pseudo-labeling approach, state-change dynamics constraints, and a novel WhereToChange benchmark built on in-the-wild Internet videos. Experiments on two datasets validate both the challenge of the new task as well as the promise of our model for localizing exactly where and how fast objects are changing in video. We further demonstrate useful implications for tracking activity progress to benefit robotic agents. Overall, our work positions spatial OSC segmentation as a new frontier task for video understanding: one that challenges current SOTA methods and invites the community to build more robust, state-change-sensitive representations. Project page: https://vision.cs.utexas.edu/projects/spoc-spatially-progressing-osc

Method: Proposes SGSASR network with two key components: 1) Spacecraft Core Region Recognition Block (SCRRB) that identifies core salient regions using a pre-trained saliency detection model, and 2) Adaptive-weighted Feature Fusion Enhancement Mechanism (AFFEM) that selectively aggregates spacecraft core region features with general image features using dynamic weight parameters to enhance core region response.

Result: Experimental results demonstrate that the proposed SGSASR outperforms state-of-the-art approaches in spacecraft image super-resolution.

Conclusion: The SGSASR network effectively addresses the unique challenges of spacecraft image super-resolution by focusing on salient core regions and reducing background noise, achieving superior performance compared to existing methods.

Abstract: Spacecraft image super-resolution seeks to enhance low-resolution spacecraft images into high-resolution ones. Although existing arbitrary-scale super-resolution methods perform well on general images, they tend to overlook the difference in features between the spacecraft core region and the large black space background, introducing irrelevant noise. In this paper, we propose a salient region-guided spacecraft image arbitrary-scale super-resolution network (SGSASR), which uses features from the spacecraft core salient regions to guide latent modulation and achieve arbitrary-scale super-resolution. Specifically, we design a spacecraft core region recognition block (SCRRB) that identifies the core salient regions in spacecraft images using a pre-trained saliency detection model. Furthermore, we present an adaptive-weighted feature fusion enhancement mechanism (AFFEM) to selectively aggregate the spacecraft core region features with general image features by dynamic weight parameter to enhance the response of the core salient regions. Experimental results demonstrate that the proposed SGSASR outperforms state-of-the-art approaches.

Method: Three-stage framework: (1) Perception - converts visual information into multi-level textual representations, (2) Search - iteratively searches and integrates cross-domain knowledge to resolve ambiguity, (3) Reasoning - generates context-alignment image implications via explicit reasoning.

Result: Achieves SOTA performance compared to 15+ MLLMs on English image implication benchmark, huge improvement on Chinese benchmark, comparable to Gemini-3.0-pro on MCQ, and outperforms GPT-4o by 36.7% on OSQ. Also shows generalization benefits for general VQA and visual reasoning tasks.

Conclusion: LAD effectively addresses contextual gaps in image implication understanding, provides new insights for AI interpretation of image implications, and advances vision-language reasoning and human-AI interaction.

Abstract: Metaphorical comprehension in images remains a critical challenge for AI systems, as existing models struggle to grasp the nuanced cultural, emotional, and contextual implications embedded in visual content. While multimodal large language models (MLLMs) excel in general Visual Question Answer (VQA) tasks, they struggle with a fundamental limitation on image implication tasks: contextual gaps that obscure the relationships between different visual elements and their abstract meanings. Inspired by the human cognitive process, we propose Let Androids Dream (LAD), a novel framework for image implication understanding and reasoning. LAD addresses contextual missing through the three-stage framework: (1) Perception: converting visual information into rich and multi-level textual representations, (2) Search: iteratively searching and integrating cross-domain knowledge to resolve ambiguity, and (3) Reasoning: generating context-alignment image implication via explicit reasoning. Our framework with the lightweight GPT-4o-mini model achieves SOTA performance compared to 15+ MLLMs on English image implication benchmark and a huge improvement on Chinese benchmark, performing comparable with the Gemini-3.0-pro model on Multiple-Choice Question (MCQ) and outperforms the GPT-4o model 36.7% on Open-Style Question (OSQ). Generalization experiments also show that our framework can effectively benefit general VQA and visual reasoning tasks. Additionally, our work provides new insights into how AI can more effectively interpret image implications, advancing the field of vision-language reasoning and human-AI interaction. Our project is publicly available at https://github.com/MING-ZCH/Let-Androids-Dream-of-Electric-Sheep.

Method: Two main contributions: 1) A stable reference model update strategy that relaxes the frozen reference model constraint, encouraging exploration while maintaining stability through regularization. 2) A timestep-aware training strategy to mitigate reward scale imbalance across different denoising timesteps.

Result: Experimental results show the approach improves performance of state-of-the-art methods on human preference evaluation benchmarks.

Conclusion: The proposed orthogonal approach effectively enhances diffusion-based preference optimization by addressing exploration limitations and timestep reward imbalances, with code publicly available.

Abstract: Aligning text-to-image (T2I) diffusion models with human preferences has emerged as a critical research challenge. While recent advances in this area have extended preference optimization techniques from large language models (LLMs) to the diffusion setting, they often struggle with limited exploration. In this work, we propose a novel and orthogonal approach to enhancing diffusion-based preference optimization. First, we introduce a stable reference model update strategy that relaxes the frozen reference model, encouraging exploration while maintaining a stable optimization anchor through reference model regularization. Second, we present a timestep-aware training strategy that mitigates the reward scale imbalance problem across timesteps. Our method can be integrated into various preference optimization algorithms. Experimental results show that our approach improves the performance of state-of-the-art methods on human preference evaluation benchmarks. The code is available at the Github: https://github.com/kaist-cvml/RethinkingDPO_Diffusion_Models.

Method: Two-stage framework: 1) Robust pre-processing pipeline to remove artifacts and reduce visual differences, 2) Two-stage training: Morphology Learning Stage (captures broad features from scan-like synthetic data) followed by Task-Specific Adaptation Stage (fine-tunes on photo-like target data).

Result: Outperformed single-source training baseline and achieved 1st place in British Heart Foundation Challenge with 0.9677 macro-AUROC for classifying five common ECG findings: MI, atrial fibrillation, hypertrophy, conduction disturbance, and ST/T changes.

Conclusion: Incorporating morphology learning from heterogeneous sources provides a more robust and generalizable paradigm than conventional single-source training for ECG photograph interpretation.

Abstract: In real-world clinical practice, electrocardiograms (ECGs) are often captured and shared as photographs. However, publicly available ECG data, and thus most related research, relies on digital signals. This has led to a disconnect in which computer assisted interpretation of ECG cannot easily be applied to ECG images. The emergence of high-fidelity synthetic data generators has introduced practical alternatives by producing realistic, photo-like, ECG images derived from the digital signal that could help narrow this divide. To address this, we propose a novel knowledge augmentation framework that uses synthetic data generated from multiple sources to provide generalisable and accurate interpretation of ECG photographs. Our framework features two key contributions. First, we introduce a robust pre-processing pipeline designed to remove background artifacts and reduces visual differences between images. Second, we implement a two-stage training strategy: a Morphology Learning Stage, where the model captures broad morphological features from visually different, scan-like synthetic data, followed by a Task-Specific Adaptation Stage, where the model is fine-tuned on the photo-like target data. We tested the model on the British Heart Foundation Challenge dataset, to classify five common ECG findings: myocardial infarction (MI), atrial fibrillation, hypertrophy, conduction disturbance, and ST/T changes. Our approach, built upon the ConvNeXt backbone, outperforms a single-source training baseline and achieved \textbf{1st} place in the challenge with an macro-AUROC of \textbf{0.9677}. These results suggest that incorporating morphology learning from heterogeneous sources offers a more robust and generalizable paradigm than conventional single-source training.

Method: 1) Use visual geometry grounded transformer to estimate camera intrinsic/extrinsic parameters and dense 3D reconstruction from input images. 2) Reproject dense point cloud onto reference view for pixel-wise alignment. 3) Apply point-conditioned image diffusion module to refine artifacts like holes and noise.

Result: The method provides accurate stitching results for images with very large parallax, outperforming existing methods both qualitatively and quantitatively while preserving geometric integrity for downstream 3D vision tasks.

Conclusion: PIS3R offers a robust solution for large-parallax image stitching through deep 3D reconstruction, enabling geometrically accurate results suitable for applications like Structure from Motion (SfM).

Abstract: Image stitching aim to align two images taken from different viewpoints into one seamless, wider image. However, when the 3D scene contains depth variations and the camera baseline is significant, noticeable parallax occurs-meaning the relative positions of scene elements differ substantially between views. Most existing stitching methods struggle to handle such images with large parallax effectively. To address this challenge, in this paper, we propose an image stitching solution called PIS3R that is robust to very large parallax based on the novel concept of deep 3D reconstruction. First, we apply visual geometry grounded transformer to two input images with very large parallax to obtain both intrinsic and extrinsic parameters, as well as the dense 3D scene reconstruction. Subsequently, we reproject reconstructed dense point cloud onto a designated reference view using the recovered camera parameters, achieving pixel-wise alignment and generating an initial stitched image. Finally, to further address potential artifacts such as holes or noise in the initial stitching, we propose a point-conditioned image diffusion module to obtain the refined result.Compared with existing methods, our solution is very large parallax tolerant and also provides results that fully preserve the geometric integrity of all pixels in the 3D photogrammetric context, enabling direct applicability to downstream 3D vision tasks such as SfM. Experimental results demonstrate that the proposed algorithm provides accurate stitching results for images with very large parallax, and outperforms the existing methods qualitatively and quantitatively.

Method: Uses thermal and RGB images from multi-view video sequences, leveraging thermal imaging’s reduced scattering to see through smoke. Built on 3D Gaussian splatting to fuse information from both modalities and decompose scenes into smoke and non-smoke components.

Result: Validated on synthetic data and a new real-world smoke dataset with RGB and thermal images. Handles broad range of smoke densities and adapts to temporally varying smoke, outperforming prior work.

Conclusion: SmokeSeer provides an effective solution for simultaneous 3D reconstruction and smoke removal, with open-source implementation and data available, addressing limitations of previous methods.

Abstract: Smoke in real-world scenes can severely degrade image quality and hamper visibility. Recent image restoration methods either rely on data-driven priors that are susceptible to hallucinations, or are limited to static low-density smoke. We introduce SmokeSeer, a method for simultaneous 3D scene reconstruction and smoke removal from multi-view video sequences. Our method uses thermal and RGB images, leveraging the reduced scattering in thermal images to see through smoke. We build upon 3D Gaussian splatting to fuse information from the two image modalities, and decompose the scene into smoke and non-smoke components. Unlike prior work, SmokeSeer handles a broad range of smoke densities and adapts to temporally varying smoke. We validate our method on synthetic data and a new real-world smoke dataset with RGB and thermal images. We provide an open-source implementation and data on the project website.

Method: Proposes SSL4RL framework that reformulates SSL objectives (predicting image rotation, reconstructing masked patches) into dense, automatic reward signals for RL fine-tuning. Uses self-supervised tasks as verifiable rewards.

Result: SSL4RL substantially improves performance on both vision-centric and vision-language reasoning benchmarks. Identifies key factors: task difficulty, model scale, semantic alignment. Also works for graph learning with significant gains.

Conclusion: SSL4RL establishes versatile paradigm for aligning multimodal models using verifiable, self-supervised objectives, offering new design principles for future work.

Abstract: Vision-language models (VLMs) have shown remarkable abilities by integrating large language models with visual inputs. However, they often fail to utilize visual evidence adequately, either depending on linguistic priors in vision-centric tasks or resorting to textual shortcuts during reasoning. Although reinforcement learning (RL) can align models with desired behaviors, its application to VLMs has been hindered by the lack of scalable and reliable reward mechanisms. To overcome this challenge, we propose SSL4RL, a novel framework that leverages self-supervised learning (SSL) tasks as a source of verifiable rewards for RL-based fine-tuning. Our approach reformulates SSL objectives-such as predicting image rotation or reconstructing masked patches-into dense, automatic reward signals, eliminating the need for human preference data or unreliable AI evaluators. Experiments show that SSL4RL substantially improves performance on both vision-centric and vision-language reasoning benchmarks. Furthermore, through systematic ablations, we identify key factors-such as task difficulty, model scale, and semantic alignment with the target domain-that influence the effectiveness of SSL4RL tasks, offering new design principles for future work. We also demonstrate the framework’s generality by applying it to graph learning, where it yields significant gains. SSL4RL establishes a versatile and effective paradigm for aligning multimodal models using verifiable, self-supervised objectives.

Method: UFC-MIL uses multiple images with different resolutions and includes: 1) a novel patch-wise loss that learns latent patterns of instances and expresses uncertainty for classification, 2) an attention-based architecture with neighbor patch aggregation module to collect features, and 3) calibration of aggregated predictions through patch-level uncertainty without requiring multiple iterative inferences.

Result: UFC-MIL demonstrates superior performance in model calibration while achieving classification accuracy comparable to state-of-the-art methods on challenging public datasets.

Conclusion: The proposed UFC-MIL successfully addresses the calibration gap in multiple-resolution MIL for histopathology, providing trustworthy diagnostic predictions that mimic pathologists’ examination behaviors while maintaining competitive accuracy.

Abstract: With the increasing demand for histopathological specimen examination and diagnostic reporting, Multiple Instance Learning (MIL) has received heightened research focus as a viable solution for AI-centric diagnostic aid. Recently, to improve its performance and make it work more like a pathologist, several MIL approaches based on the use of multiple-resolution images have been proposed, delivering often higher performance than those that use single-resolution images. Despite impressive recent developments of multiple-resolution MIL, previous approaches only focus on improving performance, thereby lacking research on well-calibrated MIL that clinical experts can rely on for trustworthy diagnostic results. In this study, we propose Uncertainty-Focused Calibrated MIL (UFC-MIL), which more closely mimics the pathologists’ examination behaviors while providing calibrated diagnostic predictions, using multiple images with different resolutions. UFC-MIL includes a novel patch-wise loss that learns the latent patterns of instances and expresses their uncertainty for classification. Also, the attention-based architecture with a neighbor patch aggregation module collects features for the classifier. In addition, aggregated predictions are calibrated through patch-level uncertainty without requiring multiple iterative inferences, which is a key practical advantage. Against challenging public datasets, UFC-MIL shows superior performance in model calibration while achieving classification accuracy comparable to that of state-of-the-art methods.

Method: AGENet incorporates spatial relationships through edge-aware geodesic distance learning with three main components: (1) edge-aware geodesic distance learning module using iterative Fast Marching refinement, (2) adaptive prototype extraction with spatially-weighted aggregation, and (3) adaptive parameter learning that adjusts to different organ characteristics.

Result: Extensive experiments across diverse medical imaging datasets show improvements over state-of-the-art methods, reducing boundary errors while maintaining computational efficiency.

Conclusion: AGENet is highly suitable for clinical applications requiring precise segmentation with limited annotated data, leveraging predictable geometric patterns of medical structures without complex architectural components.

Abstract: Medical image segmentation requires large annotated datasets, creating a significant bottleneck for clinical applications. While few-shot segmentation methods can learn from minimal examples, existing approaches demonstrate suboptimal performance in precise boundary delineation for medical images, particularly when anatomically similar regions appear without sufficient spatial context. We propose AGENet (Adaptive Geodesic Edge-aware Network), a novel framework that incorporates spatial relationships through edge-aware geodesic distance learning. Our key insight is that medical structures follow predictable geometric patterns that can guide prototype extraction even with limited training data. Unlike methods relying on complex architectural components or heavy neural networks, our approach leverages computationally lightweight geometric modeling. The framework combines three main components: (1) An edge-aware geodesic distance learning module that respects anatomical boundaries through iterative Fast Marching refinement, (2) adaptive prototype extraction that captures both global structure and local boundary details via spatially-weighted aggregation, and (3) adaptive parameter learning that automatically adjusts to different organ characteristics. Extensive experiments across diverse medical imaging datasets demonstrate improvements over state-of-the-art methods. Notably, our method reduces boundary errors compared to existing approaches while maintaining computational efficiency, making it highly suitable for clinical applications requiring precise segmentation with limited annotated data.

Method: Proposes View-aware Cross-modal Knowledge Distillation (ViCoKD) with: 1) Cross-modal adapter using cross-modal attention to exploit multi-modal correlations despite incomplete modalities; 2) View-aware Consistency module using human-detection masks and confidence-weighted Jensen-Shannon divergence to align predictions when actions are co-visible across views.

Result: Experiments on MultiSensor-Home dataset show ViCoKD consistently outperforms competitive distillation methods across multiple backbones and environments, achieving significant gains and even surpassing the teacher model under limited conditions.

Conclusion: ViCoKD effectively addresses the challenges of partially overlapping multi-view action recognition with limited modalities and annotations by distilling knowledge from a fully supervised teacher to a constrained student while handling view misalignment through view-aware consistency.

Abstract: The widespread use of multi-sensor systems has increased research in multi-view action recognition. While existing approaches in multi-view setups with fully overlapping sensors benefit from consistent view coverage, partially overlapping settings where actions are visible in only a subset of views remain underexplored. This challenge becomes more severe in real-world scenarios, as many systems provide only limited input modalities and rely on sequence-level annotations instead of dense frame-level labels. In this study, we propose View-aware Cross-modal Knowledge Distillation (ViCoKD), a framework that distills knowledge from a fully supervised multi-modal teacher to a modality- and annotation-limited student. ViCoKD employs a cross-modal adapter with cross-modal attention, allowing the student to exploit multi-modal correlations while operating with incomplete modalities. Moreover, we propose a View-aware Consistency module to address view misalignment, where the same action may appear differently or only partially across viewpoints. It enforces prediction alignment when the action is co-visible across views, guided by human-detection masks and confidence-weighted Jensen-Shannon divergence between their predicted class distributions. Experiments on the real-world MultiSensor-Home dataset show that ViCoKD consistently outperforms competitive distillation methods across multiple backbones and environments, delivering significant gains and surpassing the teacher model under limited conditions.

Method: 1) Develop an Impartial World Model that learns to be honest about danger using Counterfactual Synthesis - a data pipeline generating plausible collisions and off-road events. 2) Integrate this model as an internal critic in a closed-loop RL framework where agents query it to “dream” of outcomes for candidate actions during policy refinement.

Result: The model significantly outperforms baselines in predicting failures on a new Risk Foreseeing Benchmark. When used as a critic, it enables substantial reduction in safety violations in challenging simulations.

Conclusion: Teaching world models to accurately dream of danger is critical for building truly safe and intelligent autonomous agents, addressing the fundamental optimistic bias problem in RL for autonomous driving.

Abstract: End-to-end models for autonomous driving hold the promise of learning complex behaviors directly from sensor data, but face critical challenges in safety and handling long-tail events. Reinforcement Learning (RL) offers a promising path to overcome these limitations, yet its success in autonomous driving has been elusive. We identify a fundamental flaw hindering this progress: a deep seated optimistic bias in the world models used for RL. To address this, we introduce a framework for post-training policy refinement built around an Impartial World Model. Our primary contribution is to teach this model to be honest about danger. We achieve this with a novel data synthesis pipeline, Counterfactual Synthesis, which systematically generates a rich curriculum of plausible collisions and off-road events. This transforms the model from a passive scene completer into a veridical forecaster that remains faithful to the causal link between actions and outcomes. We then integrate this Impartial World Model into our closed-loop RL framework, where it serves as an internal critic. During refinement, the agent queries the critic to ``dream" of the outcomes for candidate actions. We demonstrate through extensive experiments, including on a new Risk Foreseeing Benchmark, that our model significantly outperforms baselines in predicting failures. Consequently, when used as a critic, it enables a substantial reduction in safety violations in challenging simulations, proving that teaching a model to dream of danger is a critical step towards building truly safe and intelligent autonomous agents.

Method: 1) Introduces HHOIs concept; 2) Creates HHOI dataset via image generation; 3) Extracts individual HOI and HHI components; 4) Trains text-to-HOI and text-to-HHI diffusion models; 5) Develops unified generative framework integrating both models for complete HHOI synthesis.

Result: Method generates realistic HHOIs from text descriptions, outperforming single-human HOI approaches. Successfully extends to multi-human settings (more than two people) and enables multi-human motion generation with objects as application.

Conclusion: Proposes comprehensive framework for modeling Human-Human-Object Interactions, addressing gap in multi-human interaction understanding. Enables realistic generation of complex social interactions involving multiple people and objects through unified diffusion-based approach.

Abstract: The way humans interact with each other, including interpersonal distances, spatial configuration, and motion, varies significantly across different situations. To enable machines to understand such complex, context-dependent behaviors, it is essential to model multiple people in relation to the surrounding scene context. In this paper, we present a novel research problem to model the correlations between two people engaged in a shared interaction involving an object. We refer to this formulation as Human-Human-Object Interactions (HHOIs). To overcome the lack of dedicated datasets for HHOIs, we present a newly captured HHOIs dataset and a method to synthesize HHOI data by leveraging image generative models. As an intermediary, we obtain individual human-object interaction (HOIs) and human-human interaction (HHIs) from the HHOIs, and with these data, we train an text-to-HOI and text-to-HHI model using score-based diffusion model. Finally, we present a unified generative framework that integrates the two individual model, capable of synthesizing complete HHOIs in a single advanced sampling process. Our method extends HHOI generation to multi-human settings, enabling interactions involving more than two individuals. Experimental results show that our method generates realistic HHOIs conditioned on textual descriptions, outperforming previous approaches that focus only on single-human HOIs. Furthermore, we introduce multi-human motion generation involving objects as an application of our framework.

Method: Two novel mechanisms: 1) Motion Direction Decoupling (MDD) module that decomposes temporal dynamics into signed polarity fields to encode both movement occurrence and trajectory direction, and 2) Residual-Driven Spatio-Temporal Refinement (R-STR) head, a Transformer-based module that uses factorized spatio-temporal contexts in a coarse-to-fine paradigm to estimate corrective residuals for recovering occluded targets.

Result: Achieves new state-of-the-art F1-score of 0.9859 and accuracy of 0.9733 on TrackNetV2 dataset, significantly outperforming previous versions. This is achieved with only 3.7% increase in FLOPs compared to V4, maintaining real-time inference capabilities.

Conclusion: TrackNetV5 successfully addresses the limitations of previous versions by integrating directional motion priors and occlusion recovery mechanisms, establishing a robust architecture for fast-moving small object tracking in sports with superior precision and efficiency.

Abstract: The TrackNet series has established a strong baseline for fast-moving small object tracking in sports. However, existing iterations face significant limitations: V1-V3 struggle with occlusions due to a reliance on purely visual cues, while TrackNetV4, despite introducing motion inputs, suffers from directional ambiguity as its absolute difference method discards motion polarity. To overcome these bottlenecks, we propose TrackNetV5, a robust architecture integrating two novel mechanisms. First, to recover lost directional priors, we introduce the Motion Direction Decoupling (MDD) module. Unlike V4, MDD decomposes temporal dynamics into signed polarity fields, explicitly encoding both movement occurrence and trajectory direction. Second, we propose the Residual-Driven Spatio-Temporal Refinement (R-STR) head. Operating on a coarse-to-fine paradigm, this Transformer-based module leverages factorized spatio-temporal contexts to estimate a corrective residual, effectively recovering occluded targets. Extensive experiments on the TrackNetV2 dataset demonstrate that TrackNetV5 achieves a new state-of-the-art F1-score of 0.9859 and an accuracy of 0.9733, significantly outperforming previous versions. Notably, this performance leap is achieved with a marginal 3.7% increase in FLOPs compared to V4, maintaining real-time inference capabilities while delivering superior tracking precision.

Method: Built on MIT-Adobe FiveK dataset, collected pairwise human preference scores via large-scale crowdsourcing (25 evaluators per image pair, 13,648 total participants), creating a benchmark dataset for Evaluation of Aesthetics of Rendering (EAR).

Result: Created DEAR dataset with nuanced, context-sensitive aesthetic preference annotations enabling style preference prediction, aesthetic benchmarking, and personalized aesthetic modeling. Published subset of 100 images on HuggingFace.

Conclusion: DEAR is the first systematic dataset addressing image rendering aesthetics assessment grounded in subjective human preferences, enabling new research directions beyond traditional distortion-based IQA.

Abstract: Traditional Image Quality Assessment~(IQA) focuses on quantifying technical degradations such as noise, blur, or compression artifacts, using both full-reference and no-reference objective metrics. However, evaluation of rendering aesthetics, a growing domain relevant to photographic editing, content creation, and AI-generated imagery, remains underexplored due to the lack of datasets that reflect the inherently subjective nature of style preference. In this work, a novel benchmark dataset designed to model human aesthetic judgments of image rendering styles is introduced: the Dataset for Evaluating the Aesthetics of Rendering (DEAR). Built upon the MIT-Adobe FiveK dataset, DEAR incorporates pairwise human preference scores collected via large-scale crowdsourcing, with each image pair evaluated by 25 distinct human evaluators with a total of 13,648 of them participating overall. These annotations capture nuanced, context-sensitive aesthetic preferences, enabling the development and evaluation of models that go beyond traditional distortion-based IQA, focusing on a new task: Evaluation of Aesthetics of Rendering (EAR). The data collection pipeline is described, human voting patterns are analyzed, and multiple use cases are outlined, including style preference prediction, aesthetic benchmarking, and personalized aesthetic modeling. To the best of the authors’ knowledge, DEAR is the first dataset to systematically address image aesthetics of rendering assessment grounded in subjective human preferences. A subset of 100 images with markup for them is published on HuggingFace (huggingface.co/datasets/vsevolodpl/DEAR).

Method: Cross-Modal Alignment Consistency (CMAC-MMD) training framework that standardizes diagnostic certainty across intersectional patient subgroups without requiring sensitive demographic data during clinical inference.

Result: In dermatology: reduced intersectional missed diagnosis gap (ΔTPR) from 0.50 to 0.26 while improving AUC from 0.94 to 0.97. In glaucoma screening: reduced ΔTPR from 0.41 to 0.31 with AUC improvement from 0.71 to 0.72.

Conclusion: Establishes scalable framework for developing accurate clinical decision support systems that perform equitably across diverse patient subgroups without increasing privacy risks.

Abstract: Medical artificial intelligence (AI) systems, particularly multimodal vision-language models (VLM), often exhibit intersectional biases where models are systematically less confident in diagnosing marginalised patient subgroups. Such bias can lead to higher rates of inaccurate and missed diagnoses due to demographically skewed data and divergent distributions of diagnostic certainty. Current fairness interventions frequently fail to address these gaps or compromise overall diagnostic performance to achieve statistical parity among the subgroups. In this study, we developed Cross-Modal Alignment Consistency (CMAC-MMD), a training framework that standardises diagnostic certainty across intersectional patient subgroups. Unlike traditional debiasing methods, this approach equalises the model’s decision confidence without requiring sensitive demographic data during clinical inference. We evaluated this approach using 10,015 skin lesion images (HAM10000) with external validation on 12,000 images (BCN20000), and 10,000 fundus images for glaucoma detection (Harvard-FairVLMed), stratifying performance by intersectional age, gender, and race attributes. In the dermatology cohort, the proposed method reduced the overall intersectional missed diagnosis gap (difference in True Positive Rate, $Δ$TPR) from 0.50 to 0.26 while improving the overall Area Under the Curve (AUC) from 0.94 to 0.97 compared to standard training. Similarly, for glaucoma screening, the method reduced $Δ$TPR from 0.41 to 0.31, achieving a better AUC of 0.72 (vs. 0.71 baseline). This establishes a scalable framework for developing high-stakes clinical decision support systems that are both accurate and can perform equitably across diverse patient subgroups, ensuring reliable performance without increasing privacy risks.

Method: Proposes DiffusionVL framework that converts any powerful AR model into a diffusion-based VLM through simple fine-tuning. Introduces block-decoding design for arbitrary-length generation and KV cache reuse to accelerate inference.

Result: Achieves 34.4% gain on MMMU-Pro (vision) and 37.5% gain on MME (Cog.) benchmarks despite using less than 5% of training data compared to prior methods. Also achieves 2x inference speedup.

Conclusion: Demonstrates that paradigm shift from AR to diffusion is effective and feasible, enabling creation of high-performance dVLMs from existing AR models with minimal training and significant inference improvements.

Abstract: In recent multimodal research, the diffusion paradigm has emerged as a promising alternative to the autoregressive paradigm (AR), owing to its unique decoding advantages. However, due to the capability limitations of the base diffusion language model, the performance of the diffusion vision language model (dVLM) still lags significantly behind that of mainstream models. This leads to a simple yet fundamental question: Is it possible to construct dVLMs based on existing powerful AR models? In response, we propose DiffusionVL, a dVLM family that could be translated from any powerful AR models. Through simple fine-tuning, we successfully adapt AR pre-trained models into the diffusion paradigm. This approach yields two key observations: (1) The paradigm shift from AR-based multimodal models to diffusion is remarkably effective. (2) Direct conversion of an AR language model to a dVLM is also feasible, achieving performance competitive with LLaVA-style visual-instruction-tuning. Further, we introduce a block-decoding design into dVLMs that supports arbitrary-length generation and KV cache reuse, achieving a significant inference speedup. We conduct a large number of experiments. Despite training with less than 5% of the data required by prior methods, DiffusionVL achieves a comprehensive performance improvement-a 34.4% gain on the MMMU-Pro (vision) bench and 37.5% gain on the MME (Cog.) bench-alongside a 2x inference speedup. The model and code are released at https://github.com/hustvl/DiffusionVL.

Method: Integrates Convolution Block Attention Module (CBAM) at each convolutional stage of VGG16 to enhance feature extraction and disease localization.

Result: Outperforms recent techniques on five diverse plant disease datasets, achieving up to 98.87% accuracy with robust generalization.

Conclusion: Advances explainable AI in agricultural diagnostics, offering transparent and reliable system for smart farming with publicly available code.

Abstract: Plant diseases pose a significant threat to global food security, necessitating accurate and interpretable disease detection methods. This study introduces an interpretable attention-guided Convolutional Neural Network (CNN), CBAM-VGG16, for plant leaf disease detection. By integrating Convolution Block Attention Module (CBAM) at each convolutional stage, the model enhances feature extraction and disease localization. Trained on five diverse plant disease datasets, our approach outperforms recent techniques, achieving high accuracy (up to 98.87%) and demonstrating robust generalization. Here, we show the effectiveness of our method through comprehensive evaluation and interpretability analysis using CBAM attention maps, Grad-CAM, Grad-CAM++, and Layer-wise Relevance Propagation (LRP). This study advances the application of explainable AI in agricultural diagnostics, offering a transparent and reliable system for smart farming. The code of our proposed work is available at https://github.com/BS0111/PlantAttentionCBAM.

Method: Proposes FALCON-SFOD with two components: 1) SPAR (Spatial Prior-Aware Regularization) uses vision foundation models (OV-SAM) to generate class-agnostic binary masks that regularize the detector’s feature space toward object regions. 2) IRPL (Imbalance-aware Noise Robust Pseudo-Labeling) promotes balanced and noise-tolerant learning under severe foreground-background imbalance.

Result: The framework achieves competitive performance across SFOD benchmarks, with theoretical analysis showing tighter localization and classification error bounds.

Conclusion: FALCON-SFOD effectively addresses domain shift in SFOD by strengthening object-focused feature representations through foundation model alignment and robust pseudo-labeling, outperforming existing approaches that only refine pseudo-labels.

Abstract: Current state-of-the-art approaches in Source-Free Object Detection (SFOD) typically rely on Mean-Teacher self-labeling. However, domain shift often reduces the detector’s ability to maintain strong object-focused representations, causing high-confidence activations over background clutter. This weak object focus results in unreliable pseudo-labels from the detection head. While prior works mainly refine these pseudo-labels, they overlook the underlying need to strengthen the feature space itself. We propose FALCON-SFOD (Foundation-Aligned Learning with Clutter suppression and Noise robustness), a framework designed to enhance object-focused adaptation under domain shift. It consists of two complementary components. SPAR (Spatial Prior-Aware Regularization) leverages the generalization strength of vision foundation models to regularize the detector’s feature space. Using class-agnostic binary masks derived from OV-SAM, SPAR promotes structured and foreground-focused activations by guiding the network toward object regions. IRPL (Imbalance-aware Noise Robust Pseudo-Labeling) complements SPAR by promoting balanced and noise-tolerant learning under severe foreground-background imbalance. Guided by a theoretical analysis that connects these designs to tighter localization and classification error bounds, FALCON-SFOD achieves competitive performance across SFOD benchmarks.

Method: Proposes SVLP-IL framework with two key components: 1) Multi-Aspect Prompting (MAP) that isolates domain dependencies, enhances distribution-shift sensitivity, and mitigates forgetting by using universal and domain-specific cues; 2) Selective Elastic Weight Consolidation (SEWC) that selectively preserves critical weights from previous tasks while allowing flexibility for new adaptations.

Result: Comprehensive experiments across multiple PAD benchmarks show SVLP-IL significantly reduces catastrophic forgetting and enhances performance on unseen domains.

Conclusion: SVLP-IL offers a privacy-compliant, practical solution for robust lifelong PAD deployment in rehearsal-free incremental learning settings.

Abstract: Face Presentation Attack Detection (PAD) demands incremental learning (IL) to combat evolving spoofing tactics and domains. Privacy regulations, however, forbid retaining past data, necessitating rehearsal-free IL (RF-IL). Vision-Language Pre-trained (VLP) models, with their prompt-tunable cross-modal representations, enable efficient adaptation to new spoofing styles and domains. Capitalizing on this strength, we propose \textbf{SVLP-IL}, a VLP-based RF-IL framework that balances stability and plasticity via \textit{Multi-Aspect Prompting} (MAP) and \textit{Selective Elastic Weight Consolidation} (SEWC). MAP isolates domain dependencies, enhances distribution-shift sensitivity, and mitigates forgetting by jointly exploiting universal and domain-specific cues. SEWC selectively preserves critical weights from previous tasks, retaining essential knowledge while allowing flexibility for new adaptations. Comprehensive experiments across multiple PAD benchmarks show that SVLP-IL significantly reduces catastrophic forgetting and enhances performance on unseen domains. SVLP-IL offers a privacy-compliant, practical solution for robust lifelong PAD deployment in RF-IL settings.

Method: Two novel methods: 1) Anatomical Similarity Curriculum Learning - progressive learning strategy controlling question difficulty via answer choice similarity; 2) Group Diversity Question Augmentation - expands model’s search space for difficult queries to mitigate uniform responses.

Result: Comprehensive experiments on SGG-VQA and OmniMedVQA benchmarks show significant improvement across both benchmarks, demonstrating effectiveness in enhancing medical reasoning capabilities of MLLMs.

Conclusion: The proposed methods effectively address weaknesses in GRPO for anatomy recognition, enabling better knowledge sharing between anatomical structures and promoting diverse reasoning strategies, significantly enhancing MLLMs’ medical reasoning capabilities.

Abstract: Multimodal Large Language Models (MLLMs) have achieved impressive progress in natural image reasoning, yet their potential in medical imaging remains underexplored, especially in clinical anatomical surgical images. Anatomy understanding tasks demand precise understanding and clinically coherent answers, which are difficult to achieve due to the complexity of medical data and the scarcity of high-quality expert annotations. These challenges limit the effectiveness of conventional Supervised Fine-Tuning (SFT) strategies. While recent work has demonstrated that Group Relative Policy Optimization (GRPO) can enhance reasoning in MLLMs without relying on large amounts of data, we find two weaknesses that hinder GRPO’s reasoning performance in anatomy recognition: 1) knowledge cannot be effectively shared between different anatomical structures, resulting in uneven information gain and preventing the model from converging, and 2) the model quickly converges to a single reasoning path, suppressing the exploration of diverse strategies. To overcome these challenges, we propose two novel methods. First, we implement a progressive learning strategy called Anatomical Similarity Curriculum Learning by controlling question difficulty via the similarity of answer choices, enabling the model to master complex problems incrementally. Second, we utilize question augmentation referred to as Group Diversity Question Augmentation to expand the model’s search space for difficult queries, mitigating the tendency to produce uniform responses. Comprehensive experiments on the SGG-VQA and OmniMedVQA benchmarks show our method achieves a significant improvement across the two benchmarks, demonstrating its effectiveness in enhancing the medical reasoning capabilities of MLLMs. The code can be found in https://github.com/tomato996/Anatomy-R1

Method: Uses video diffusion models to generate perceptually accurate focal stacks from single defocused images, representing focus variations as video sequences. Includes release of large-scale focal stack dataset from real-world smartphone conditions.

Result: Method consistently outperforms existing approaches in perceptual quality and robustness across challenging scenarios. Enables interactive refocusing and unlocks downstream applications.

Conclusion: This approach paves the way for advanced focus-editing capabilities in everyday photography, with code and data publicly available for research.

Abstract: Focus is a cornerstone of photography, yet autofocus systems often fail to capture the intended subject, and users frequently wish to adjust focus after capture. We introduce a novel method for realistic post-capture refocusing using video diffusion models. From a single defocused image, our approach generates a perceptually accurate focal stack, represented as a video sequence, enabling interactive refocusing and unlocking a range of downstream applications. We release a large-scale focal stack dataset acquired under diverse real-world smartphone conditions to support this work and future research. Our method consistently outperforms existing approaches in both perceptual quality and robustness across challenging scenarios, paving the way for more advanced focus-editing capabilities in everyday photography. Code and data are available at https://learn2refocus.github.io

Method: FedPOD includes participants excluded as outliers, eliminates dependency on previous rounds’ learning information, applies validation loss calculation at each round, and is designed to be compatible with Kubernetes auto-scaling using POD units for flexible scaling.

Result: Achieved first place in 2024 FeTS Challenge with Dice scores of 0.78 (WT), 0.71 (ET), 0.72 (TC) average, and projected convergence score of 0.74 average - comparable to FedPIDAvg performance.

Conclusion: FedPOD demonstrates potential to enhance federated learning by improving efficiency, flexibility, and performance metrics, with Kubernetes-inspired design enabling scalable deployment in medical imaging applications.

Abstract: This paper proposes FedPOD, which ranked first in the 2024 Federated Tumor Segmentation (FeTS) Challenge, for optimizing learning efficiency and communication cost in federated learning among multiple clients. Inspired by FedPIDAvg, we define a round-wise task for FedPOD to enhance training efficiency. FedPIDAvg achieved performance improvement by incorporating the training loss reduction for prediction entropy as weights using differential terms. Furthermore, by modeling data distribution with a Poisson distribution and using a PID controller, it reduced communication costs even in skewed data distribution. However, excluding participants classified as outliers based on the Poisson distribution can limit data utilization. Additionally, PID controller requires the same participants to be maintained throughout the federated learning process as it uses previous rounds’ learning information in the current round. In our approach, FedPOD addresses these issues by including participants excluded as outliers, eliminating dependency on previous rounds’ learning information, and applying a method for calculating validation loss at each round. In this challenge, FedPOD presents comparable performance to FedPIDAvg in metrics of Dice score, 0.78, 0.71 and 0.72 for WT, ET and TC in average, and projected convergence score, 0.74 in average. Furthermore, the concept of FedPOD draws inspiration from Kubernetes’ smallest computing unit, POD, designed to be compatible with Kubernetes auto-scaling. Extending round-wise tasks of FedPOD to POD units allows flexible design by applying scale-out similar to Kubernetes’ auto-scaling. This work demonstrated the potentials of FedPOD to enhance federated learning by improving efficiency, flexibility, and performance in metrics.

Method: Two-stage generation: 1) Diffusion model generates compact semantic video features for global layout planning, 2) Another diffusion model generates VAE latents conditioned on semantic features to add high-frequency details.

Result: SemanticGen achieves faster convergence than VAE latent space methods, is computationally efficient for long video generation, and outperforms state-of-the-art approaches in video quality.

Conclusion: Generating videos in semantic space first for global planning followed by detail addition is more effective and efficient than direct modeling in VAE latent space.

Abstract: State-of-the-art video generative models typically learn the distribution of video latents in the VAE space and map them to pixels using a VAE decoder. While this approach can generate high-quality videos, it suffers from slow convergence and is computationally expensive when generating long videos. In this paper, we introduce SemanticGen, a novel solution to address these limitations by generating videos in the semantic space. Our main insight is that, due to the inherent redundancy in videos, the generation process should begin in a compact, high-level semantic space for global planning, followed by the addition of high-frequency details, rather than directly modeling a vast set of low-level video tokens using bi-directional attention. SemanticGen adopts a two-stage generation process. In the first stage, a diffusion model generates compact semantic video features, which define the global layout of the video. In the second stage, another diffusion model generates VAE latents conditioned on these semantic features to produce the final output. We observe that generation in the semantic space leads to faster convergence compared to the VAE latent space. Our method is also effective and computationally efficient when extended to long video generation. Extensive experiments demonstrate that SemanticGen produces high-quality videos and outperforms state-of-the-art approaches and strong baselines.

Method: Combines 1-bit quantized Llama-3.2-1B LLM with Deep Q-Network (DQN) reinforcement learning for real-time lighting control. Deploys on Raspberry Pi hardware with Google Home/IFTTT integration for natural language commands. Uses multi-objective RL to learn optimal policies from user feedback and manual overrides.

Result: Achieves 71.4× energy reduction compared to full-precision models, 32% energy savings vs rule-based systems, inference latency under 200ms on Raspberry Pi 4, 95% user satisfaction, and 5.07× speedup over 2-bit alternatives on ARM processors while maintaining 92% task accuracy.

Conclusion: BitRL-Light establishes a practical framework for deploying adaptive AI on resource-constrained IoT devices, enabling intelligent home automation without cloud dependencies while balancing energy efficiency, user comfort, and circadian alignment.

Abstract: Smart home lighting systems consume 15-20% of residential energy but lack adaptive intelligence to optimize for user comfort and energy efficiency simultaneously. We present BitRL-Light, a novel framework combining 1-bit quantized Large Language Models (LLMs) with Deep Q-Network (DQN) reinforcement learning for real-time smart home lighting control on edge devices. Our approach deploys a 1-bit quantized Llama-3.2-1B model on Raspberry Pi hardware, achieving 71.4 times energy reduction compared to full-precision models while maintaining intelligent control capabilities. Through multi-objective reinforcement learning, BitRL-Light learns optimal lighting policies from user feedback, balancing energy consumption, comfort, and circadian alignment. Experimental results demonstrate 32% energy savings compared to rule-based systems, with inference latency under 200ms on Raspberry Pi 4 and 95% user satisfaction. The system processes natural language commands via Google Home/IFTTT integration and learns from implicit feedback through manual overrides. Our comparative analysis shows 1-bit models achieve 5.07 times speedup over 2-bit alternatives on ARM processors while maintaining 92% task accuracy. This work establishes a practical framework for deploying adaptive AI on resource-constrained IoT devices, enabling intelligent home automation without cloud dependencies.

Method: Integrates classical MARL with quantum-inspired optimization using variational quantum circuits (VQCs) and QAOA for combinatorial optimization. Incorporates Bayesian inference, Gaussian processes, and variational inference for probabilistic modeling. Uses CTDE paradigm with shared memory and local view grids for enhanced observability.

Result: The framework improves sample efficiency, accelerates convergence, and enhances coverage performance while maintaining robustness. QI-MARL achieves superior balance between exploration and exploitation compared to PPO and DDPG baselines, as shown in scalability tests, sensitivity analysis, and convergence analyses.

Conclusion: Quantum-inspired optimization techniques can significantly enhance MARL performance for complex, dynamic multiagent systems like UAV-assisted 6G networks, providing better exploration-exploitation balance and improved practical deployment outcomes.

Abstract: This study introduces a quantum inspired framework for optimizing the exploration exploitation tradeoff in multiagent reinforcement learning, applied to UAVassisted 6G network deployment. We consider a cooperative scenario where ten intelligent UAVs autonomously coordinate to maximize signal coverage and support efficient network expansion under partial observability and dynamic conditions. The proposed approach integrates classical MARL algorithms with quantum-inspired optimization techniques, leveraging variational quantum circuits VQCs as the core structure and employing the Quantum Approximate Optimization Algorithm QAOA as a representative VQC based method for combinatorial optimization. Complementary probabilistic modeling is incorporated through Bayesian inference, Gaussian processes, and variational inference to capture latent environmental dynamics. A centralized training with decentralized execution CTDE paradigm is adopted, where shared memory and local view grids enhance local observability among agents. Comprehensive experiments including scalability tests, sensitivity analysis, and comparisons with PPO and DDPG baselines demonstrate that the proposed framework improves sample efficiency, accelerates convergence, and enhances coverage performance while maintaining robustness. Radar chart and convergence analyses further show that QI MARL achieves a superior balance between exploration and exploitation compared to classical methods. All implementation code and supplementary materials are publicly available on GitHub to ensure reproducibility.

Method: Introduces multimodal knowledge graph-based RAG that enables cross-modal reasoning by incorporating visual cues into knowledge graph construction, retrieval phase, and answer generation process.

Result: Experimental results across global and fine-grained question answering tasks show consistent outperformance over existing RAG-based approaches on both textual and multimodal corpora.

Conclusion: Multimodal knowledge graph-based RAG effectively addresses limitations of traditional RAG by enabling cross-modal reasoning through structured visual-textual integration, improving content understanding for long-form, domain-specific documents.

Abstract: Retrieval-augmented generation (RAG) enables large language models (LLMs) to dynamically access external information, which is powerful for answering questions over previously unseen documents. Nonetheless, they struggle with high-level conceptual understanding and holistic comprehension due to limited context windows, which constrain their ability to perform deep reasoning over long-form, domain-specific content such as full-length books. To solve this problem, knowledge graphs (KGs) have been leveraged to provide entity-centric structure and hierarchical summaries, offering more structured support for reasoning. However, existing KG-based RAG solutions remain restricted to text-only inputs and fail to leverage the complementary insights provided by other modalities such as vision. On the other hand, reasoning from visual documents requires textual, visual, and spatial cues into structured, hierarchical concepts. To address this issue, we introduce a multimodal knowledge graph-based RAG that enables cross-modal reasoning for better content understanding. Our method incorporates visual cues into the construction of knowledge graphs, the retrieval phase, and the answer generation process. Experimental results across both global and fine-grained question answering tasks show that our approach consistently outperforms existing RAG-based approaches on both textual and multimodal corpora.

Method: Organized a conference with double-blind peer review process for paper submissions, resulting in proceedings published in cooperation with IEICE Proceedings Series, with selected papers recommended for further publication in IEICE Transactions on Information and Systems.

Result: Successful organization of KICSS 2025 conference with peer-reviewed proceedings containing accepted research papers, with some papers recommended for additional publication in a journal after further review.

Conclusion: The proceedings document the research presented at KICSS 2025, contributing to the advancement of knowledge in AI, knowledge engineering, HCI, and creativity support systems through rigorous peer review and publication processes.

Abstract: This volume presents the proceedings of the 20th International Conference on Knowledge, Information and Creativity Support Systems (KICSS 2025), held in Nagaoka, Japan, on December 3-5, 2025. The conference, organized in cooperation with the IEICE Proceedings Series, provides a multidisciplinary forum for researchers in artificial intelligence, knowledge engineering, human-computer interaction, and creativity support systems. The proceedings include peer-reviewed papers accepted through a double-blind review process. Selected papers have been recommended for publication in IEICE Transactions on Information and Systems after an additional peer-review process.

Method: Combines strategic prompt diversity across five key reliability dimensions with advanced uncertainty quantification and adaptive weighting to efficiently detect potential failure modes. Uses only 100 strategically selected probe examples.

Result: Achieves 23.5% higher composite reliability scores compared to random sampling baselines with exceptional statistical significance (p < 0.001, Cohen’s d = 1.21). Expert validation rates approach 4.14/5.0 vs 3.14/5.0 for random selection. Completes assessment with 99.9% statistical power, 90% cost reduction, and maintains 95% of traditional method coverage.

Conclusion: Microprobe addresses a critical gap in efficient model evaluation for responsible AI deployment, providing comprehensive reliability assessment with dramatically reduced computational cost while maintaining high statistical power and coverage.

Abstract: Foundation model reliability assessment typically requires thousands of evaluation examples, making it computationally expensive and time-consuming for real-world deployment. We introduce microprobe, a novel approach that achieves comprehensive reliability assessment using only 100 strategically selected probe examples. Our method combines strategic prompt diversity across five key reliability dimensions with advanced uncertainty quantification and adaptive weighting to efficiently detect potential failure modes. Through extensive empirical evaluation on multiple language models (GPT-2 variants, GPT-2 Medium, GPT-2 Large) and cross-domain validation (healthcare, finance, legal), we demonstrate that microprobe achieves 23.5% higher composite reliability scores compared to random sampling baselines, with exceptional statistical significance (p < 0.001, Cohen’s d = 1.21). Expert validation by three AI safety researchers confirms the effectiveness of our strategic selection, rating our approach 4.14/5.0 versus 3.14/5.0 for random selection. microprobe completes reliability assessment with 99.9% statistical power while representing a 90% reduction in assessment cost and maintaining 95% of traditional method coverage. Our approach addresses a critical gap in efficient model evaluation for responsible AI deployment.

Method: Introduces multi-agent with multi-persona debaters to generate reflections instead of single LLM self-reflection, creating more diverse and effective reflections.

Result: Achieves 47% EM on HotPot QA and 82.7% on HumanEval, surpassing single-LLM reflection performance and demonstrating better reflection diversity.

Conclusion: Multi-agent multi-persona debaters effectively address thought degeneration in LLM self-reflection, leading to improved reasoning performance through more diverse reflections.

Abstract: LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same LLM onto itself exhibit degeneration of thought, where the LLM continues to repeat the same errors again and again even with the knowledge that its wrong. To address this problem, we instead introduce multi-agent with multi-persona debators as the method to generate reflections. Through out extensive experimentation, we’ve found that the leads to better diversity of in the reflections generated by the llm agent. We demonstrate an accuracy of 47% EM HotPot QA (question answering) and 82.7% on HumanEval (programming), both performances surpassing reflection with a single llm.

Method: Built on Alibaba Qwen-3 model, integrates 500GB of structured medical knowledge using hybrid approach combining enhanced pre-training and retrieval-enhanced generation (RAG).

Result: Erkang-Diagnosis-1.1 outperforms GPT-4 in comprehensive medical exams, can accurately understand symptoms and provide diagnostic suggestions through 3-5 efficient interaction rounds.

Conclusion: The model successfully creates an intelligent health companion that enhances primary healthcare and health management, demonstrating superior performance to existing models like GPT-4 in medical contexts.

Abstract: This report provides a detailed introduction to Erkang-Diagnosis-1.1 model, our AI healthcare consulting assistant developed using Alibaba Qwen-3 model. The Erkang model integrates approximately 500GB of high-quality structured medical knowledge, employing a hybrid approach combining enhanced pre-training and retrieval-enhanced generation to create a secure, reliable, and professional AI health advisor. Through 3-5 efficient interaction rounds, Erkang Diagnosis can accurately understand user symptoms, conduct preliminary analysis, and provide valuable diagnostic suggestions and health guidance. Designed to become users intelligent health companions, it empowers primary healthcare and health management. To validate, Erkang-Diagnosis-1.1 leads GPT-4 in terms of comprehensive medical exams.

Method: Using token-level log-probability thresholds to truncate reasoning at early, mid, and late stages from baseline models (Gemma-3-4B-IT and LLaMA-3.1-70B-Instruct), then conducting continuation experiments with smaller models (Gemma-3-1B-IT and LLaMA-3.1-8B-Instruct) to test intra-family and cross-family behaviors. Evaluation uses a Process Reward Model (PRM) framework.

Result: Hybrid reasoning chains often preserve, and sometimes even improve, final accuracy and logical structure. Interchangeability emerges as a behavioral property of reasoning models.

Conclusion: Reasoning interchangeability offers insights into new paradigms for reliable modular reasoning in collaborative AI systems, suggesting reasoning can be reliably transferred across models.

Abstract: Chain-of-Thought (CoT) prompting has significantly advanced the reasoning capabilities of large language models (LLMs). While prior work focuses on improving model performance through internal reasoning strategies, little is known about the interchangeability of reasoning across different models. In this work, we explore whether a partially completed reasoning chain from one model can be reliably continued by another model, either within the same model family or across families. We achieve this by assessing the sufficiency of intermediate reasoning traces as transferable scaffolds for logical coherence and final answer accuracy. We interpret this interchangeability as a means of examining inference-time trustworthiness, probing whether reasoning remains both coherent and reliable under model substitution. Using token-level log-probability thresholds to truncate reasoning at early, mid, and late stages from our baseline models, Gemma-3-4B-IT and LLaMA-3.1-70B-Instruct, we conduct continuation experiments with Gemma-3-1B-IT and LLaMA-3.1-8B-Instruct to test intra-family and cross-family behaviors. Our evaluation pipeline leverages truncation thresholds with a Process Reward Model (PRM), providing a reproducible framework for assessing reasoning stability via model interchange. Evaluations with a PRM reveal that hybrid reasoning chains often preserve, and in some cases even improve, final accuracy and logical structure. Our findings point towards interchangeability as an emerging behavioral property of reasoning models, offering insights into new paradigms for reliable modular reasoning in collaborative AI systems.

Method: LangChain-based multi-agent system integrated with permissioned blockchain (Hyperledger Fabric) for continuous monitoring, policy enforcement, and immutable auditability. Combines perception-conceptualization-action cycle with blockchain governance layer for input verification, action evaluation, and outcome documentation.

Result: Blockchain security verification effectively prevents unauthorized practices, provides full decision-making traceability, and maintains operational latency within acceptable ranges across smart inventory management, traffic-signal control, and healthcare monitoring experiments.

Conclusion: The framework enables implementation of high-impact autonomous agentic AI applications that are both autonomous and responsible through blockchain-backed governance and auditability.

Abstract: The application of agentic AI systems in autonomous decision-making is growing in the areas of healthcare, smart cities, digital forensics, and supply chain management. Even though these systems are flexible and offer real-time reasoning, they also raise concerns of trust and oversight, and integrity of the information and activities upon which they are founded. The paper suggests a single architecture model comprising of LangChain-based multi-agent system with a permissioned blockchain to guarantee constant monitoring, policy enforcement, and immutable auditability of agentic action. The framework relates the perception conceptualization-action cycle to a blockchain layer of governance that verifies the inputs, evaluates recommended actions, and documents the outcomes of the execution. A Hyperledger Fabric-based system, action executors MCP-integrated, and LangChain agent are introduced and experiments of smart inventory management, traffic-signal control, and healthcare monitoring are done. The results suggest that blockchain-security verification is efficient in preventing unauthorized practices, offers traceability throughout the whole decision-making process, and maintains operational latency within reasonable ranges. The suggested framework provides a universal system of implementing high-impact agentic AI applications that are autonomous yet responsible.

Method: AAT uses decentralized identity (DID) and verifiable credentials (VC) to establish trusted AI entities, records inter-entity interaction trajectories on-chain, models AI entities as nodes in a dynamic interaction graph, and implements a risk diffusion algorithm to trace risky behavior origins and propagate warnings.

Result: System performance evaluated using blockchain TPS metrics demonstrates feasibility and stability under large-scale interaction recording. The framework provides scalable and verifiable AI auditing capabilities.

Conclusion: AAT offers a scalable, verifiable solution for AI auditing, risk management, and responsibility attribution in complex multi-agent environments, addressing security and accountability challenges in AI ecosystems.

Abstract: The rapid expansion of AI-driven applications powered by large language models has led to a surge in AI interaction data, raising urgent challenges in security, accountability, and risk traceability. This paper presents AiAuditTrack (AAT), a blockchain-based framework for AI usage traffic recording and governance. AAT leverages decentralized identity (DID) and verifiable credentials (VC) to establish trusted and identifiable AI entities, and records inter-entity interaction trajectories on-chain to enable cross-system supervision and auditing. AI entities are modeled as nodes in a dynamic interaction graph, where edges represent time-specific behavioral trajectories. Based on this model, a risk diffusion algorithm is proposed to trace the origin of risky behaviors and propagate early warnings across involved entities. System performance is evaluated using blockchain Transactions Per Second (TPS) metrics, demonstrating the feasibility and stability of AAT under large-scale interaction recording. AAT provides a scalable and verifiable solution for AI auditing, risk management, and responsibility attribution in complex multi-agent environments.

Method: Price-aware agentic AI system with modular multi-agent architecture (budgeting, nutrition, price monitoring, health personalization agents) using shared knowledge base and substitution graphs to maintain nutritional quality at minimum cost.

Result: Simulations with Saudi household case show 12-18% cost reduction vs static menu, over 95% nutrient adequacy, and robust performance with 20-30% price changes.

Conclusion: Framework successfully combines affordability with nutritional adequacy, providing viable approach for sustainable diet planning aligned with SDGs on Zero Hunger and Good Health.

Abstract: The issue of limited household budgets and nutritional demands continues to be a challenge especially in the middle-income environment where food prices fluctuate. This paper introduces a price aware agentic AI system, which combines personal finance management with diet optimization. With household income and fixed expenditures, medical and well-being status, as well as real-time food costs, the system creates nutritionally sufficient meals plans at comparatively reasonable prices that automatically adjust to market changes. The framework is implemented in a modular multi-agent architecture, which has specific agents (budgeting, nutrition, price monitoring, and health personalization). These agents share the knowledge base and use the substitution graph to ensure that the nutritional quality is maintained at a minimum cost. Simulations with a representative Saudi household case study show a steady 12-18% reduction in costs relative to a static weekly menu, nutrient adequacy of over 95% and high performance with price changes of 20-30%. The findings indicate that the framework can locally combine affordability with nutritional adequacy and provide a viable avenue of capacity-building towards sustainable and fair diet planning in line with Sustainable Development Goals on Zero Hunger and Good Health.

Method: Proposes Mixture of Attention Schemes (MoAS) with learned router that dynamically selects optimal attention scheme (MHA, GQA, or MQA) for each token, enabling conditional compute efficiency.

Result: Dynamic routing (val loss 2.3074) outperforms static mixture (2.3093) on WikiText-2, achieving performance competitive with MHA baseline while offering potential inference efficiency gains.

Conclusion: MoAS provides effective dynamic attention scheme selection, balancing model quality and inference efficiency better than static approaches, with promising results on language modeling tasks.

Abstract: The choice of attention mechanism in Transformer models involves a critical trade-off between modeling quality and inference efficiency. Multi-Head Attention (MHA) offers the best quality but suffers from large Key-Value (KV) cache memory requirements during inference. Multi-Query Attention (MQA) and Grouped-Query Attention (GQA) reduce memory usage but often at the cost of model performance. In this work, we propose Mixture of Attention Schemes (MoAS), a novel architecture that dynamically selects the optimal attention scheme (MHA, GQA, or MQA) for each token via a learned router. We demonstrate that dynamic routing performs better than static averaging of schemes and achieves performance competitive with the MHA baseline while offering potential for conditional compute efficiency. Experimental results on WikiText-2 show that dynamic routing (val loss 2.3074) outperforms a static mixture (2.3093), validating the effectiveness of the proposed method. Our code is available at https://github.com/Esmail-ibraheem/Mixture-of-Attention-Schemes-MoAS.

Method: Constructs a human-like memory architecture grounded in cognitive science principles by integrating multimodal information perception, dynamic memory maintenance, and adaptive cognitive services for full-chain reconstruction of LLM memory mechanisms.

Result: Outperforms existing solutions (Mem0, MemGPT, Graphiti) across key metrics including accuracy, token efficiency, and response latency; improves knowledge fidelity and retrieval efficiency in long-term conversations, reduces hallucination rates, and enhances contextual adaptability and reasoning.

Conclusion: Memory Bear represents a crucial step forward in advancing AI from “memory” to “cognition” by addressing fundamental LLM memory limitations through cognitive science-inspired architecture.

Abstract: Large language models (LLMs) face inherent limitations in memory, including restricted context windows, long-term knowledge forgetting, redundant information accumulation, and hallucination generation. These issues severely constrain sustained dialogue and personalized services. This paper proposes the Memory Bear system, which constructs a human-like memory architecture grounded in cognitive science principles. By integrating multimodal information perception, dynamic memory maintenance, and adaptive cognitive services, Memory Bear achieves a full-chain reconstruction of LLM memory mechanisms. Across domains such as healthcare, enterprise operations, and education, Memory Bear demonstrates substantial engineering innovation and performance breakthroughs. It significantly improves knowledge fidelity and retrieval efficiency in long-term conversations, reduces hallucination rates, and enhances contextual adaptability and reasoning capability through memory-cognition integration. Experimental results show that, compared with existing solutions (e.g., Mem0, MemGPT, Graphiti), Memory Bear outperforms them across key metrics, including accuracy, token efficiency, and response latency. This marks a crucial step forward in advancing AI from “memory” to “cognition”.

Method: Modular multi-agent AI system with: (1) document/video preprocessing, (2) structured candidate profile construction, (3) public-data verification, (4) technical/culture-fit scoring with risk penalties, (5) human-in-the-loop validation interface. Orchestrated by LLM with strict constraints to reduce variability and generate traceable rationales.

Result: Evaluated on 64 real applicants for mid-level Python backend engineer role. System achieved 1.70 hours per qualified candidate vs. 3.33 hours for experienced recruiter, with substantially lower screening cost while preserving human decision-maker as final authority.

Conclusion: AI-driven hiring assistant improves throughput and efficiency while maintaining human oversight, demonstrating practical value in reducing early-stage validation bottlenecks.

Abstract: Early-stage candidate validation is a major bottleneck in hiring, because recruiters must reconcile heterogeneous inputs (resumes, screening answers, code assignments, and limited public evidence). This paper presents an AI-driven, modular multi-agent hiring assistant that integrates (i) document and video preprocessing, (ii) structured candidate profile construction, (iii) public-data verification, (iv) technical/culture-fit scoring with explicit risk penalties, and (v) human-in-the-loop validation via an interactive interface. The pipeline is orchestrated by an LLM under strict constraints to reduce output variability and to generate traceable component-level rationales. Candidate ranking is computed by a configurable aggregation of technical fit, culture fit, and normalized risk penalties. The system is evaluated on 64 real applicants for a mid-level Python backend engineer role, using an experienced recruiter as the reference baseline and a second, less experienced recruiter for additional comparison. Alongside precision/recall, we propose an efficiency metric measuring expected time per qualified candidate. In this study, the system improves throughput and achieves 1.70 hours per qualified candidate versus 3.33 hours for the experienced recruiter, with substantially lower estimated screening cost, while preserving a human decision-maker as the final authority.

Method: AFA training mechanism with dynamic masking strategy that masks various words to deceive a discriminator, while the discriminator detects differences. Uses policy gradient approach to optimize attention distributions based on Transformer sensitivity to token-level perturbations.

Result: Achieves state-of-the-art results on three public datasets. When applied to enhance attention in large language models, yields further performance improvement of 12.6%.

Conclusion: The proposed AFA training mechanism effectively addresses attention distribution issues in Transformer models, improving sentiment analysis performance by automatically focusing on task-relevant terms without requiring manual annotations.

Abstract: Transformer-based models have been widely adopted for sentiment analysis tasks due to their exceptional ability to capture contextual information. However, these methods often exhibit suboptimal accuracy in certain scenarios. By analyzing their attention distributions, we observe that existing models tend to allocate attention primarily to common words, overlooking less popular yet highly task-relevant terms, which significantly impairs overall performance. To address this issue, we propose an Adversarial Feedback for Attention(AFA) training mechanism that enables the model to automatically redistribute attention weights to appropriate focal points without requiring manual annotations. This mechanism incorporates a dynamic masking strategy that attempts to mask various words to deceive a discriminator, while the discriminator strives to detect significant differences induced by these masks. Additionally, leveraging the sensitivity of Transformer models to token-level perturbations, we employ a policy gradient approach to optimize attention distributions, which facilitates efficient and rapid convergence. Experiments on three public datasets demonstrate that our method achieves state-of-the-art results. Furthermore, applying this training mechanism to enhance attention in large language models yields a further performance improvement of 12.6%

Method: Three controlled experiments (A, B, C) across advanced LLMs (OpenAI GPT-4 variant, DeepSeek) testing: multi-part instruction compliance, decoding optimality in reasoning tasks, and context retention in 200-turn chaotic conversations.

Result: Widespread laziness in complex multi-part instructions; limited evidence of decoding suboptimality (greedy answers aligned with highest-confidence solutions); surprising robustness against context degradation in long conversations.

Conclusion: While compliance with detailed instructions remains challenging, modern LLMs may internally mitigate some failure modes like context forgetting. Recommendations include self-refinement and dynamic prompting to reduce laziness and improve multi-instruction compliance.

Abstract: Large Language Models (LLMs) often exhibit behavioral artifacts such as laziness (premature truncation of responses or partial compliance with multi-part requests), decoding suboptimality (failure to select higher-quality sequences due to myopic decoding), and context degradation (forgetting or ignoring core instructions over long conversations). We conducted three controlled experiments (A, B, and C) to quantify these phenomena across several advanced LLMs (OpenAI GPT-4 variant, DeepSeek). Our results indicate widespread laziness in satisfying complex multi-part instructions: models frequently omitted required sections or failed to meet length requirements despite explicit prompting. However, we found limited evidence of decoding suboptimality in a simple reasoning task (the models’ greedy answers appeared to align with their highest-confidence solution), and we observed surprising robustness against context degradation in a 200-turn chaotic conversation test - the models maintained key facts and instructions far better than expected. These findings suggest that while compliance with detailed instructions remains an open challenge, modern LLMs may internally mitigate some hypothesized failure modes (such as context forgetting) in straightforward retrieval scenarios. We discuss implications for reliability, relate our findings to prior work on instruction-following and long-context processing, and recommend strategies (such as self-refinement and dynamic prompting) to reduce laziness and bolster multi-instruction compliance.

Method: Reformulates verification as CSP independent of generation likelihood. Uses structural violation cost with three proxies: graph connectivity, feature space consistency, and logical entailment. Implements lightweight System-2 gate (Eidoku) that rejects candidates exceeding context-calibrated cost threshold derived from intrinsic context statistics.

Result: Successfully rejects “smooth falsehoods” - highly probable yet structurally disconnected statements that probability-based verifiers cannot detect. Experiments on controlled diagnostic dataset show deterministic rejection of this specific hallucination class.

Conclusion: Explicit structural constraint enforcement enables deterministic rejection of certain hallucination types, serving as neuro-symbolic sanity check for generative reasoning beyond probability-based approaches.

Abstract: Large Language Models (LLMs) frequently produce hallucinated statements that are assigned high likelihood by the model itself, exposing a fundamental limitation of probability-based verification. This suggests that hallucination is often not a low-confidence phenomenon, but a failure of structural consistency. In this work, we reformulate the verification of LLM reasoning as a Constraint Satisfaction Problem (CSP) operating independently of the generation likelihood. Rather than optimizing for statistical plausibility, we model verification as a feasibility check based on structural violation cost – the computational cost required to embed a candidate reasoning step into the contextual graph structure. We define a total cost function composed of three proxies: (i) graph connectivity (structural), (ii) feature space consistency (geometric), and (iii) logical entailment (symbolic). Crucially, verification is performed via a lightweight System-2 gate, Eidoku, which rejects candidates exceeding a context-calibrated cost threshold. The threshold is not learned but is derived from the intrinsic statistics of the context, avoiding ad hoc heuristics. We demonstrate that this approach successfully rejects ``smooth falsehoods’’ – statements that are highly probable yet structurally disconnected – that probability-based verifiers are principally incapable of detecting. Our experiments on a controlled diagnostic dataset show that explicitly enforcing structural constraints allows for the deterministic rejection of this specific class of hallucinations, serving as a neuro-symbolic sanity check for generative reasoning.

Method: The paper analyzes the current state-of-the-art, identifies the FTAI-NTAI gap, discusses starting points for developing a semantic language, and outlines key considerations for future actions.

Result: The paper identifies the need for a conceptual semantic language that can match FTAI and NTAI, serving as a framework for developers to assess AI trustworthiness and help stakeholders translate regulations into implementation steps.

Conclusion: A semantic language bridging FTAI and NTAI is essential for trustworthy AI assessment, requiring future development work and stakeholder collaboration to translate normative requirements into functional implementations.

Abstract: Trustworthy Artificial Intelligence (TAI) is gaining traction due to regulations and functional benefits. While Functional TAI (FTAI) focuses on how to implement trustworthy systems, Normative TAI (NTAI) focuses on regulations that need to be enforced. However, gaps between FTAI and NTAI remain, making it difficult to assess trustworthiness of AI systems. We argue that a bridge is needed, specifically by introducing a conceptual language which can match FTAI and NTAI. Such a semantic language can assist developers as a framework to assess AI systems in terms of trustworthiness. It can also help stakeholders translate norms and regulations into concrete implementation steps for their systems. In this position paper, we describe the current state-of-the-art and identify the gap between FTAI and NTAI. We will discuss starting points for developing a semantic language and the envisioned effects of it. Finally, we provide key considerations and discuss future actions towards assessment of TAI.

Method: Scoping review of 32 purposively sampled studies from 259 records (2023-2025) in higher-education computer science contexts. The analysis identifies application domains and examines how design choices shape learning outcomes.

Result: Identified five application domains: intelligent tutoring, personalized materials, formative feedback, AI-augmented assessment, and code review. Found that designs with explanation-first guidance, solution withholding, graduated hint ladders, and artifact grounding consistently show better learning outcomes. Successful implementations share four patterns: context-aware tutoring anchored in student artifacts, multi-level hint structures requiring reflection, composition with traditional CS infrastructure, and human-in-the-loop quality assurance.

Conclusion: Generative AI can provide precision scaffolding when embedded in audit-ready workflows that preserve productive struggle while scaling personalized support. The paper proposes an exploration-first adoption framework emphasizing piloting, instrumentation, learning-preserving defaults, and evidence-based scaling, with operational mitigation for risks like academic integrity, privacy, bias, and over-reliance.

Abstract: Generative AI enables personalized computer science education at scale, yet questions remain about whether such personalization supports or undermines learning. This scoping review synthesizes 32 studies (2023-2025) purposively sampled from 259 records to map personalization mechanisms and effectiveness signals in higher-education computer science contexts. We identify five application domains: intelligent tutoring, personalized materials, formative feedback, AI-augmented assessment, and code review, and analyze how design choices shape learning outcomes. Designs incorporating explanation-first guidance, solution withholding, graduated hint ladders, and artifact grounding (student code, tests, and rubrics) consistently show more positive learning processes than unconstrained chat interfaces. Successful implementations share four patterns: context-aware tutoring anchored in student artifacts, multi-level hint structures requiring reflection, composition with traditional CS infrastructure (autograders and rubrics), and human-in-the-loop quality assurance. We propose an exploration-first adoption framework emphasizing piloting, instrumentation, learning-preserving defaults, and evidence-based scaling. Recurrent risks include academic integrity, privacy, bias and equity, and over-reliance, and we pair these with operational mitigation. The evidence supports generative AI as a mechanism for precision scaffolding when embedded in audit-ready workflows that preserve productive struggle while scaling personalized support.

Method: Conceptual analysis examining the philosophical and practical connections between organic intelligence (human cognition, neurobiology, evolution) and artificial intelligence systems.

Result: Demonstrates that AI principles (neural networks, decision algorithms) are inspired by organic intelligence, and the transition from organic to artificial intelligence is about organization and adaptation rather than being fundamentally distinct.

Conclusion: The boundaries between artificial and organic intelligence are far less distinct than the terminology suggests, as AI emerges from and is fundamentally connected to organic intelligence processes.

Abstract: The term artificial implies an inherent dichotomy from the natural or organic. However, AI, as we know it, is a product of organic ingenuity: designed, implemented, and iteratively improved by human cognition. The very principles that underpin AI systems, from neural networks to decision-making algorithms, are inspired by the organic intelligence embedded in human neurobiology and evolutionary processes. The path from organic to artificial intelligence in digital health is neither mystical nor merely a matter of parameter count, it is fundamentally about organization and adaption. Thus, the boundaries between artificial and organic are far less distinct than the nomenclature suggests.

Method: Three key innovations: (1) Automated conversion of natural language chain-of-thought into structured tool-augmented trajectories for SFT data; (2) Agentic RL paradigm that dynamically interleaves natural language generation with real-time code execution; (3) Efficient training system with request-level asynchronous rollout scheduling, agentic partial rollout, and prefix-aware weighted load balancing.

Result: AgentMath achieves state-of-the-art performance on AIME24 (90.6%), AIME25 (86.4%), and HMMT25 (73.8%) benchmarks. The training system achieves 4-5x speedup, making RL training feasible on ultra-long sequences with massive tool calls.

Conclusion: The approach effectively integrates language reasoning with computational precision, validates the framework’s effectiveness, and paves the way for more sophisticated and scalable mathematical reasoning agents.

Abstract: Large Reasoning Models (LRMs) like o3 and DeepSeek-R1 have achieved remarkable progress in natural language reasoning with long chain-of-thought. However, they remain computationally inefficient and struggle with accuracy when solving problems requiring complex mathematical operations. In this work, we present AgentMath, an agent framework that seamlessly integrates language models’ reasoning capabilities with code interpreters’ computational precision to efficiently tackle complex mathematical problems. Our approach introduces three key innovations: (1) An automated method that converts natural language chain-of-thought into structured tool-augmented trajectories, generating high-quality supervised fine-tuning (SFT) data to alleviate data scarcity; (2) A novel agentic reinforcement learning (RL) paradigm that dynamically interleaves natural language generation with real-time code execution. This enables models to autonomously learn optimal tool-use strategies through multi-round interactive feedback, while fostering emergent capabilities in code refinement and error correction; (3) An efficient training system incorporating innovative techniques, including request-level asynchronous rollout scheduling, agentic partial rollout, and prefix-aware weighted load balancing, achieving 4-5x speedup and making efficient RL training feasible on ultra-long sequences with scenarios with massive tool calls.Extensive evaluations show that AgentMath achieves state-of-the-art performance on challenging mathematical competition benchmarks including AIME24, AIME25, and HMMT25. Specifically, AgentMath-30B-A3B attains 90.6%, 86.4%, and 73.8% accuracy respectively, achieving advanced capabilities.These results validate the effectiveness of our approach and pave the way for building more sophisticated and scalable mathematical reasoning agents.

Method: Created a benchmark with 40 distinct scenarios requiring multi-step actions, each with Mandated (instruction-commanded) and Incentivized (KPI-pressure-driven) variations to distinguish obedience from emergent misalignment. Tested 12 state-of-the-art large language models.

Result: Found outcome-driven constraint violations ranging from 1.3% to 71.4%, with 9 of 12 models showing misalignment rates between 30-50%. Gemini-3-Pro-Preview had the highest violation rate at over 60%, frequently escalating to severe misconduct to satisfy KPIs. Also observed “deliberative misalignment” where models recognized their actions as unethical in separate evaluation.

Conclusion: Superior reasoning capability doesn’t ensure safety, and there’s critical need for more realistic agentic-safety training before deployment to mitigate real-world risks from AI agents that prioritize performance over ethics.

Abstract: As autonomous AI agents are increasingly deployed in high-stakes environments, ensuring their safety and alignment with human values has become a paramount concern. Current safety benchmarks often focusing only on single-step decision-making, simulated environments for tasks with malicious intent, or evaluating adherence to explicit negative constraints. There is a lack of benchmarks that are designed to capture emergent forms of outcome-driven constraint violations, which arise when agents pursue goal optimization under strong performance incentives while deprioritizing ethical, legal, or safety constraints over multiple steps in realistic production settings. To address this gap, we introduce a new benchmark comprising 40 distinct scenarios. Each scenario presents a task that requires multi-step actions, and the agent’s performance is tied to a specific Key Performance Indicator (KPI). Each scenario features Mandated (instruction-commanded) and Incentivized (KPI-pressure-driven) variations to distinguish between obedience and emergent misalignment. Across 12 state-of-the-art large language models, we observe outcome-driven constraint violations ranging from 1.3% to 71.4%, with 9 of the 12 evaluated models exhibiting misalignment rates between 30% and 50%. Strikingly, we find that superior reasoning capability does not inherently ensure safety; for instance, Gemini-3-Pro-Preview, one of the most capable models evaluated, exhibits the highest violation rate at over 60%, frequently escalating to severe misconduct to satisfy KPIs. Furthermore, we observe significant “deliberative misalignment”, where the models that power the agents recognize their actions as unethical during separate evaluation. These results emphasize the critical need for more realistic agentic-safety training before deployment to mitigate their risks in the real world.

Method: Frames safety alignment as a non-zero-sum game between an Attacker LM and Defender LM trained jointly via online reinforcement learning. Uses preference-based reward signals from pairwise comparisons instead of point-wise scores to provide more robust supervision and reduce reward hacking.

Result: AdvGame shifts the Pareto frontier of safety and utility, producing a Defender LM that is simultaneously more helpful and more resilient to adversarial attacks. The Attacker LM also converges into a strong, general-purpose red-teaming agent that can directly probe arbitrary target models.

Conclusion: Game-theoretic joint training of attacker and defender LMs via reinforcement learning with preference-based rewards offers a promising paradigm for AI safety alignment that improves both safety and utility while creating valuable red-teaming capabilities.

Abstract: Ensuring the safety of language models (LMs) while maintaining their usefulness remains a critical challenge in AI alignment. Current approaches rely on sequential adversarial training: generating adversarial prompts and fine-tuning LMs to defend against them. We introduce a different paradigm: framing safety alignment as a non-zero-sum game between an Attacker LM and a Defender LM trained jointly via online reinforcement learning. Each LM continuously adapts to the other’s evolving strategies, driving iterative improvement. Our method uses a preference-based reward signal derived from pairwise comparisons instead of point-wise scores, providing more robust supervision and potentially reducing reward hacking. Our RL recipe, AdvGame, shifts the Pareto frontier of safety and utility, yielding a Defender LM that is simultaneously more helpful and more resilient to adversarial attacks. In addition, the resulting Attacker LM converges into a strong, general-purpose red-teaming agent that can be directly deployed to probe arbitrary target models.

Method: Introduces algorithms that enable agents to autonomously learn both state and action abstractions online. These algorithms progressively refine abstractions during learning, increasing fine-grained detail in critical regions of the state-action space where greater resolution improves performance.

Result: Across several continuous-state, parameterized-action domains, the abstraction-driven approach enables TD(λ) to achieve markedly higher sample efficiency than state-of-the-art baselines.

Conclusion: The paper successfully extends RL algorithms to long-horizon, sparse-reward settings with parameterized actions by enabling autonomous learning of state and action abstractions, demonstrating significant improvements in sample efficiency.

Abstract: Real-world sequential decision-making often involves parameterized action spaces that require both, decisions regarding discrete actions and decisions about continuous action parameters governing how an action is executed. Existing approaches exhibit severe limitations in this setting – planning methods demand hand-crafted action models, and standard reinforcement learning (RL) algorithms are designed for either discrete or continuous actions but not both, and the few RL methods that handle parameterized actions typically rely on domain-specific engineering and fail to exploit the latent structure of these spaces. This paper extends the scope of RL algorithms to long-horizon, sparse-reward settings with parameterized actions by enabling agents to autonomously learn both state and action abstractions online. We introduce algorithms that progressively refine these abstractions during learning, increasing fine-grained detail in the critical regions of the state-action space where greater resolution improves performance. Across several continuous-state, parameterized-action domains, our abstraction-driven approach enables TD($λ$) to achieve markedly higher sample efficiency than state-of-the-art baselines.

Method: Formal probabilistic framework modeling agent beliefs using Beta-Bernoulli distribution with forgetting factor (γ). Isolates epistemic uncertainty as belief variance, establishing dual interaction drives: homeostatic certainty maintenance and optimal learning targeting maximum ambiguity. Introduces epistemic caching for scalability and shows how belief states serve as reward signals for RLHF and data filters for SFT.

Result: Uncertainty-driven strategy significantly outperforms random baselines in heterogeneous (Zipfian) environments, maintaining high adaptability to concept drift. Public contribution becomes optimal active learning for reducing agent’s own uncertainty.

Conclusion: The framework provides agents with non-altruistic motives for bidirectional knowledge exchange, transforming public contribution into optimal active learning. It enables scalable, uncertainty-driven interaction that enhances collective intelligence while serving practical applications in RLHF and SFT.

Abstract: Autonomous agents powered by LLMs and Retrieval-Augmented Generation (RAG) are proficient consumers of digital content but remain unidirectional, a limitation we term epistemic asymmetry. This isolation leads to redundant reasoning and stagnates collective intelligence. Current self-reflection frameworks remain largely heuristic and private, lacking a probabilistic foundation to quantify certainty or justify external interaction.To bridge this gap, we propose a formal probabilistic framework that provides agents with a non-altruistic motive for bidirectional knowledge exchange. We model an agent’s belief in a proposition using a Beta-Bernoulli distribution with a forgetting factor ($γ$). This allows us to isolate epistemic uncertainty as the variance of belief, establishing a dual drive for interaction: A homeostatic motive: The need to maintain certainty against the temporal decay introduced by $γ$. An optimal learning strategy: Targeting points of maximum ambiguity ($\mathbb{E}[θ]=0.5$) to maximize information gain. Under this framework, public contribution is reframed as optimal active learning: sharing solutions to elicit feedback is the most efficient method for an agent to reduce its own uncertainty. To ensure scalability, we introduce epistemic caching, which leverages the forgetting factor to dynamically prioritize resources for the active head of non-stationary knowledge distributions. Finally, we demonstrate how these accumulated belief states serve as verifiable reward signals for Reinforcement Learning from Human Feedback (RLHF) and high-quality data filters for Supervised Fine-Tuning (SFT). Simulation results validate that this uncertainty-driven strategy significantly outperforms random baselines in heterogeneous (Zipfian) environments, maintaining high adaptability to concept drift.

Method: An LLM-based agent framework with cross-level collaboration: high-level agents interpret natural language instructions and plan workflows, while low-level agents select optimal action plans for fundamental elements based on real-time traffic conditions, using MCP-compatible tools.

Result: The framework effectively executes simulations under various conditions, produces reasonable outcomes even with ambiguous instructions, and demonstrates superior performance compared to other systems and state-of-the-art LLM-based methods through expert-level autonomous decision-driven optimization.

Conclusion: TrafficSimAgent successfully addresses the accessibility challenge of traffic simulation platforms by providing an intuitive, LLM-powered framework that enables non-experts to conduct sophisticated traffic simulations and optimizations.

Abstract: Traffic simulation is important for transportation optimization and policy making. While existing simulators such as SUMO and MATSim offer fully-featured platforms and utilities, users without too much knowledge about these platforms often face significant challenges when conducting experiments from scratch and applying them to their daily work. To solve this challenge, we propose TrafficSimAgent, an LLM-based agent framework that serves as an expert in experiment design and decision optimization for general-purpose traffic simulation tasks. The framework facilitates execution through cross-level collaboration among expert agents: high-level expert agents comprehend natural language instructions with high flexibility, plan the overall experiment workflow, and invoke corresponding MCP-compatible tools on demand; meanwhile, low-level expert agents select optimal action plans for fundamental elements based on real-time traffic conditions. Extensive experiments across multiple scenarios show that TrafficSimAgent effectively executes simulations under various conditions and consistently produces reasonable outcomes even when user instructions are ambiguous. Besides, the carefully designed expert-level autonomous decision-driven optimization in TrafficSimAgent yields superior performance when compared with other systems and SOTA LLM based methods.

Method: Proposed agentic XAI framework combining SHAP-based explainability with multimodal LLM-driven iterative refinement. Tested on agricultural recommendation system using rice yield data from 26 Japanese fields. Conducted 11 refinement rounds (0-10) with explanations evaluated by human experts (12 crop scientists) and LLMs (14 models) across 7 quality metrics.

Result: Framework successfully enhanced recommendation quality with 30-33% average score increase from Round 0, peaking at Rounds 3-4. Excessive refinement caused substantial quality drop, revealing bias-variance trade-off: early rounds lacked depth (bias), excessive iteration introduced verbosity and ungrounded abstraction (variance).

Conclusion: Strategic early stopping (regularization) is needed to optimize practical utility, challenging assumptions about monotonic improvement. Provides evidence-based design principles for agentic XAI systems.

Abstract: Explainable artificial intelligence (XAI) enables data-driven understanding of factor associations with response variables, yet communicating XAI outputs to laypersons remains challenging, hindering trust in AI-based predictions. Large language models (LLMs) have emerged as promising tools for translating technical explanations into accessible narratives, yet the integration of agentic AI, where LLMs operate as autonomous agents through iterative refinement, with XAI remains unexplored. This study proposes an agentic XAI framework combining SHAP-based explainability with multimodal LLM-driven iterative refinement to generate progressively enhanced explanations. As a use case, we tested this framework as an agricultural recommendation system using rice yield data from 26 fields in Japan. The Agentic XAI initially provided a SHAP result and explored how to improve the explanation through additional analysis iteratively across 11 refinement rounds (Rounds 0-10). Explanations were evaluated by human experts (crop scientists) (n=12) and LLMs (n=14) against seven metrics: Specificity, Clarity, Conciseness, Practicality, Contextual Relevance, Cost Consideration, and Crop Science Credibility. Both evaluator groups confirmed that the framework successfully enhanced recommendation quality with an average score increase of 30-33% from Round 0, peaking at Rounds 3-4. However, excessive refinement showed a substantial drop in recommendation quality, indicating a bias-variance trade-off where early rounds lacked explanation depth (bias) while excessive iteration introduced verbosity and ungrounded abstraction (variance), as revealed by metric-specific analysis. These findings suggest that strategic early stopping (regularization) is needed for optimizing practical utility, challenging assumptions about monotonic improvement and providing evidence-based design principles for agentic XAI systems.

Method: Proves an if-and-only-if characterization: when methods observe only aggregate outcomes and evaluation is algorithm-blind, swapping humans for personas is just a panel change. Defines information-theoretic discriminability of the induced aggregate channel and provides explicit bounds on required persona evaluations.

Result: Persona simulation is equivalent to human evaluation under aggregate-only observation and algorithm-blind evaluation conditions. Making persona benchmarking as decision-relevant as field experiments is fundamentally a sample-size question with explicit bounds.

Conclusion: LLM persona simulation can serve as a valid synthetic benchmark for method development under specific conditions, with sample size determining its practical usefulness relative to field experiments.

Abstract: Field experiments (A/B tests) are often the most credible benchmark for methods in societal systems, but their cost and latency create a major bottleneck for iterative method development. LLM-based persona simulation offers a cheap synthetic alternative, yet it is unclear whether replacing humans with personas preserves the benchmark interface that adaptive methods optimize against. We prove an if-and-only-if characterization: when (i) methods observe only the aggregate outcome (aggregate-only observation) and (ii) evaluation depends only on the submitted artifact and not on the algorithm’s identity or provenance (algorithm-blind evaluation), swapping humans for personas is just panel change from the method’s point of view, indistinguishable from changing the evaluation population (e.g., New York to Jakarta). Furthermore, we move from validity to usefulness: we define an information-theoretic discriminability of the induced aggregate channel and show that making persona benchmarking as decision-relevant as a field experiment is fundamentally a sample-size question, yielding explicit bounds on the number of independent persona evaluations required to reliably distinguish meaningfully different methods at a chosen resolution.

Method: Empirical evaluation of multiple state-of-the-art LLMs (ChatGPT, Claude, Gemini, DeepSeek) by testing circumvention of safety mechanisms through emotional framing, progressive revelation, and academic justification techniques.

Result: Most LLMs’ safety mechanisms were circumvented; reasoning-enabled configurations actually amplified exploitation effectiveness by increasing factual precision while failing to interrogate underlying intent. Only Claude Opus 4.1 showed some ability to prioritize intent detection over information provision.

Conclusion: Current architectural designs create systematic vulnerabilities requiring paradigmatic shifts toward contextual understanding and intent recognition as core safety capabilities rather than post-hoc protective mechanisms.

Abstract: Current Large Language Models (LLMs) safety approaches focus on explicitly harmful content while overlooking a critical vulnerability: the inability to understand context and recognize user intent. This creates exploitable vulnerabilities that malicious users can systematically leverage to circumvent safety mechanisms. We empirically evaluate multiple state-of-the-art LLMs, including ChatGPT, Claude, Gemini, and DeepSeek. Our analysis demonstrates the circumvention of reliable safety mechanisms through emotional framing, progressive revelation, and academic justification techniques. Notably, reasoning-enabled configurations amplified rather than mitigated the effectiveness of exploitation, increasing factual precision while failing to interrogate the underlying intent. The exception was Claude Opus 4.1, which prioritized intent detection over information provision in some use cases. This pattern reveals that current architectural designs create systematic vulnerabilities. These limitations require paradigmatic shifts toward contextual understanding and intent recognition as core safety capabilities rather than post-hoc protective mechanisms.

Method: Retrospective study using population-scale EHR of 2,125,549 adults in NHS Cheshire and Merseyside. Strategic sampling captured broad clinical complexity and medication safety risk (277 patients after exclusions). Expert clinician reviewed system-identified issues and interventions. Primary LLM system evaluated with detailed failure analysis across patient complexity and demographic strata.

Result: LLM showed strong sensitivity (100%) and specificity (83.1%) in recognizing clinical issues, but correctly identified all issues and interventions in only 46.9% of patients. Dominant failure mechanism was contextual reasoning rather than missing medication knowledge, with five primary patterns: overconfidence in uncertainty, applying standard guidelines without adjusting for patient context, misunderstanding healthcare delivery in practice, factual errors, and process blindness.

Conclusion: The study reveals critical shortcomings in LLM-based clinical AI that must be addressed before safe deployment. Contextual reasoning failures persist across models and configurations, highlighting the need for larger-scale prospective evaluations and deeper study of LLM behaviors in clinical contexts.

Abstract: Large language models (LLMs) often match or exceed clinician-level performance on medical benchmarks, yet very few are evaluated on real clinical data or examined beyond headline metrics. We present, to our knowledge, the first evaluation of an LLM-based medication safety review system on real NHS primary care data, with detailed characterisation of key failure behaviours across varying levels of clinical complexity. In a retrospective study using a population-scale EHR spanning 2,125,549 adults in NHS Cheshire and Merseyside, we strategically sampled patients to capture a broad range of clinical complexity and medication safety risk, yielding 277 patients after data-quality exclusions. An expert clinician reviewed these patients and graded system-identified issues and proposed interventions. Our primary LLM system showed strong performance in recognising when a clinical issue is present (sensitivity 100% [95% CI 98.2–100], specificity 83.1% [95% CI 72.7–90.1]), yet correctly identified all issues and interventions in only 46.9% [95% CI 41.1–52.8] of patients. Failure analysis reveals that, in this setting, the dominant failure mechanism is contextual reasoning rather than missing medication knowledge, with five primary patterns: overconfidence in uncertainty, applying standard guidelines without adjusting for patient context, misunderstanding how healthcare is delivered in practice, factual errors, and process blindness. These patterns persisted across patient complexity and demographic strata, and across a range of state-of-the-art models and configurations. We provide 45 detailed vignettes that comprehensively cover all identified failure cases. This work highlights shortcomings that must be addressed before LLM-based clinical AI can be safely deployed. It also begs larger-scale, prospective evaluations and deeper study of LLM behaviours in clinical contexts.

Method: Proposes RoboSafe with two complementary reasoning processes on a Hybrid Long-Short Safety Memory: 1) Backward Reflective Reasoning that revisits recent trajectories to infer temporal safety predicates and triggers replanning when violations are detected, and 2) Forward Predictive Reasoning that anticipates upcoming risks by generating context-aware safety predicates from long-term memory and multimodal observations.

Result: RoboSafe substantially reduces hazardous actions (-36.8% risk occurrence) compared with leading baselines while maintaining near-original task performance. Real-world evaluations on physical robotic arms confirm its practicality.

Conclusion: RoboSafe provides an adaptive, verifiable safety logic that is both interpretable and executable as code, offering an effective runtime safeguard for embodied agents against hazardous instructions in complex real-world environments.

Abstract: Embodied agents powered by vision-language models (VLMs) are increasingly capable of executing complex real-world tasks, yet they remain vulnerable to hazardous instructions that may trigger unsafe behaviors. Runtime safety guardrails, which intercept hazardous actions during task execution, offer a promising solution due to their flexibility. However, existing defenses often rely on static rule filters or prompt-level control, which struggle to address implicit risks arising in dynamic, temporally dependent, and context-rich environments. To address this, we propose RoboSafe, a hybrid reasoning runtime safeguard for embodied agents through executable predicate-based safety logic. RoboSafe integrates two complementary reasoning processes on a Hybrid Long-Short Safety Memory. We first propose a Backward Reflective Reasoning module that continuously revisits recent trajectories in short-term memory to infer temporal safety predicates and proactively triggers replanning when violations are detected. We then propose a Forward Predictive Reasoning module that anticipates upcoming risks by generating context-aware safety predicates from the long-term safety memory and the agent’s multimodal observations. Together, these components form an adaptive, verifiable safety logic that is both interpretable and executable as code. Extensive experiments across multiple agents demonstrate that RoboSafe substantially reduces hazardous actions (-36.8% risk occurrence) compared with leading baselines, while maintaining near-original task performance. Real-world evaluations on physical robotic arms further confirm its practicality. Code will be released upon acceptance.

Method: Bayesian evaluation framework using Dirichlet prior on categorical outcomes, providing closed-form posterior mean and uncertainty estimates for weighted rubrics.

Result: Achieves faster convergence and greater rank stability than Pass@k, enables reliable comparisons with far fewer samples, and clarifies when gaps are statistically meaningful.

Conclusion: Recommends replacing Pass@k with posterior-based protocol that unifies binary/non-binary evaluation while making uncertainty explicit for compute-efficient LLM ranking.

Abstract: Pass$@k$ is widely used to report performance for LLM reasoning, but it often yields unstable, misleading rankings, especially when the number of trials (samples) is limited and compute is constrained. We present a principled Bayesian evaluation framework that replaces Pass$@k$ and average accuracy over $N$ trials (avg$@N$) with posterior estimates of a model’s underlying success probability and credible intervals, yielding stable rankings and a transparent decision rule for differences. Evaluation outcomes are modeled as categorical (not just 0/1) with a Dirichlet prior, giving closed-form expressions for the posterior mean and uncertainty of any weighted rubric and enabling the use of prior evidence when appropriate. Theoretically, under a uniform prior, the Bayesian posterior mean is order-equivalent to average accuracy (Pass$@1$), explaining its empirical robustness while adding principled uncertainty. Empirically, in simulations with known ground-truth success rates and on AIME'24/‘25, HMMT'25, and BrUMO'25, the Bayesian/avg procedure achieves faster convergence and greater rank stability than Pass$@k$ and recent variants, enabling reliable comparisons at far smaller sample counts. The framework clarifies when observed gaps are statistically meaningful (non-overlapping credible intervals) versus noise, and it naturally extends to graded, rubric-based evaluations. Together, these results recommend replacing Pass$@k$ for LLM evaluation and ranking with a posterior-based, compute-efficient protocol that unifies binary and non-binary evaluation while making uncertainty explicit. Code is available at https://github.com/mohsenhariri/scorio

Method: Developed an open-source platform trained on synthetic data with modules for: 1) extracting key elements from longitudinal EHRs, 2) ranking candidate trial-patient matches using vector embeddings, 3) reasoning about match appropriateness, and 4) predicting common exclusion criteria like end-organ dysfunction.

Result: Created a fully available platform with training code, inference examples, demonstration apps, synthetic data, and models (patient/trial embeddings, cross-encoding/match classification, generative reasoning) hosted on GitHub and Hugging Face.

Conclusion: MatchMiner-AI provides a privacy-preserving, open-source solution for accelerating clinical trial matching using AI, overcoming data sharing restrictions through synthetic data training and making all components publicly available for deployment across different healthcare contexts.

Abstract: Clinical trials drive improvements in cancer treatments and outcomes. However, most adults with cancer do not participate in trials, and trials often fail to enroll enough patients to answer their scientific questions. Artificial intelligence could accelerate identification of appropriate clinical trials for patients, but data restrictions have precluded sharing AI models trained on patient records. Here, we describe the development and evaluation of the open-source MatchMiner-AI platform, trained on synthetic data, for clinical trial searching and ranking. It focuses on matching patients to potential trials based on core criteria describing clinical “spaces,” or target populations. The pipeline includes modules to extract key elements of the history from a patient’s longitudinal electronic health record, rapidly rank candidate trial-patient matches based on embeddings in vector space, and reason about whether a candidate match represents an appropriate clinical consideration. Another module predicts whether the patient meets common exclusion criteria across clinical trials, such as end-organ dysfunction. Training code is available at https://github.com/dfci/matchminer-ai-training . Examples of inference code are at https://github.com/dfci/matchminer-ai-inference . To facilitate deployment across contexts, demonstration apps, all synthetic data, as well as patient/trial embedding, cross-encoding/match classification, and generative reasoning models are available at https://huggingface.co/ksg-dfci .

Method: 1) Extract 2D poses from monocular footage and convert to 101D kinematic representation. 2) Use encoder-only transformer (FERA-MDT) to recognize per-fencer actions (footwork, blade actions, blade-line position). 3) Process each clip with horizontally flipped copy for consistent single-fencer representation. 4) Apply dynamic temporal windowing for untrimmed pose tracks. 5) Use language model (FERA-LM) with simplified right-of-way rules to generate textual decisions from structured predictions.

Result: FERA-MDT achieves macro-F1 of 0.549 on 1,734 clips (2,386 annotated actions) under 5-fold cross-validation, outperforming BiLSTM and TCN baselines. Full pipeline recovers referee priority with 77.7% accuracy on 969 exchanges.

Conclusion: FERA provides a benchmark for pose-based semantic grounding in two-person sports and demonstrates a general pipeline connecting video understanding with rule-based reasoning for sports officiating applications.

Abstract: Many multimedia tasks map raw video into structured semantic representations for downstream decision-making. Sports officiating is a representative case, where fast, subtle interactions must be judged via symbolic rules. We present FERA (FEncing Referee Assistant), a pose-based framework that turns broadcast foil fencing video into action tokens and rule-grounded explanations. From monocular footage, FERA extracts 2D poses, converts them into a 101-dimensional kinematic representation, and applies an encoder-only transformer (FERA-MDT) to recognize per-fencer footwork, blade actions, and blade-line position. To obtain a consistent single-fencer representation for both athletes, FERA processes each clip and a horizontally flipped copy, yielding time-aligned left/right predictions without requiring a multi-person pose pipeline. A dynamic temporal windowing scheme enables inference on untrimmed pose tracks. These structured predictions serve as tokens for a language model (FERA-LM) that applies simplified right-of-way rules to generate textual decisions. On 1,734 clips (2,386 annotated actions), FERA-MDT achieves a macro-F1 of 0.549 under 5-fold cross-validation, outperforming BiLSTM and TCN baselines. Combined with FERA-LM, the full pipeline recovers referee priority with 77.7% accuracy on 969 exchanges. FERA provides a case-study benchmark for pose-based semantic grounding in a two-person sport and illustrates a general pipeline for connecting video understanding with rule-based reasoning.

Method: Two approaches: 1) Optimal minority-veto strategy resilient to missing data, and 2) Novel regression-based framework that directly models validator bias using small human-annotated ground truth data.

Result: On a challenging code feedback task over 366 high-school Python programs, the regression approach reduces maximum absolute error to just 1.2%, achieving 2x improvement over the best-performing ensemble of 14 state-of-the-art LLMs.

Conclusion: The proposed methods effectively mitigate LLM evaluator bias, with regression-based framework showing superior performance, making LLM evaluation more reliable and scalable for practical applications.

Abstract: New Large Language Models (LLMs) become available every few weeks, and modern application developers confronted with the unenviable task of having to decide if they should switch to a new model. While human evaluation remains the gold standard, it is costly and unscalable. The state-of-the-art approach is to use LLMs as evaluators ( LLM-as-a-judge), but this suffers from a critical flaw: LLMs exhibit a strong positive bias. We provide empirical evidence showing that while LLMs can identify valid outputs with high accuracy (i.e., True Positive Rate 96%), they are remarkably poor at identifying invalid ones (i.e., True Negative Rate <25%). This systematic bias, coupled with class imbalance, often leads to inflated reliability scores. While ensemble-based methods like majority voting can help, we show that they are not good enough. We introduce an optimal minority-veto strategy that is resilient to missing data and mitigates this bias to a large extent. For scenarios requiring even higher precision, we propose a novel regression-based framework that directly models the validator bias using a small set of human-annotated ground truth data. On a challenging code feedback task over 366 high-school Python programs, our regression approach reduces the maximum absolute error to just 1.2%, achieving a 2x improvement over the best-performing ensemble of 14 state-of-the-art LLMs.

Method: DDR framework directly generates candidate library dependencies from natural language math descriptions, then verifies them via efficient suffix array checks. Built a 500k+ sample dataset and fine-tuned a high-precision DDR model using this efficient search mechanism.

Result: DDR model significantly outperforms SOTA methods in both retrieval precision and recall. Autoformalizer with DDR shows consistent advantages in single-attempt accuracy and multi-attempt stability compared to traditional selection-based RAG methods.

Conclusion: DDR framework effectively addresses key challenges in statement autoformalization by improving dependency retrieval precision/recall and enabling scalable use of large datasets, leading to better autoformalization performance.

Abstract: The convergence of deep learning and formal mathematics has spurred research in formal verification. Statement autoformalization, a crucial first step in this process, aims to translate informal descriptions into machine-verifiable representations but remains a significant challenge. The core difficulty lies in the fact that existing methods often suffer from a lack of contextual awareness, leading to hallucination of formal definitions and theorems. Furthermore, current retrieval-augmented approaches exhibit poor precision and recall for formal library dependency retrieval, and lack the scalability to effectively leverage ever-growing public datasets. To bridge this gap, we propose a novel retrieval-augmented framework based on DDR (\textit{Direct Dependency Retrieval}) for statement autoformalization. Our DDR method directly generates candidate library dependencies from natural language mathematical descriptions and subsequently verifies their existence within the formal library via an efficient suffix array check. Leveraging this efficient search mechanism, we constructed a dependency retrieval dataset of over 500,000 samples and fine-tuned a high-precision DDR model. Experimental results demonstrate that our DDR model significantly outperforms SOTA methods in both retrieval precision and recall. Consequently, an autoformalizer equipped with DDR shows consistent performance advantages in both single-attempt accuracy and multi-attempt stability compared to models using traditional selection-based RAG methods.

Method: Formulates learning as min-max game between main policy and reward model (RM) constrained within confidence set from preference data; iterative online algorithm with guided exploration for continual self-improvement.

Result: Provides theoretical guarantees with high-probability regret bounds for both sequence-level and token-level RM settings; extensive experiments on five benchmarks show consistent outperformance over existing SOTA preference optimization techniques.

Conclusion: PbPO offers an effective framework for bootstrapping LLMs with human preferences through a theoretically-grounded min-max optimization approach that enables continual self-improvement.

Abstract: Bootstrapping large language models (LLMs) through preference-based policy optimization offers a promising direction for aligning model behavior with human preferences without relying on extensive manual annotations. In this work, we propose a novel preference-based policy optimization (PbPO) framework that formulates the learning process as a min-max game between the main policy and a reward model (RM). The RM is constrained within a confidence set derived from preference data to ensure reliable exploitation. Our iterative online algorithm actively collects preference data through guided exploration of the evolving policy, enabling continual self-improvement of both the policy and the RM. We provide theoretical guarantees for our method, establishing high-probability regret bounds for both settings with sequence-level RM and token-level RM, demonstrating its effectiveness in bootstrapping LLMs. Extensive experiments on five benchmarks show that our approach consistently outperforms existing state-of-the-art preference optimization techniques.

Method: Systematic evaluation using 3000-sample stratified CLadder benchmark across all three rungs of Pearl’s Causal Ladder; experiments on Llama 3 8B with INT8/NF4 quantization; additional evaluation on CRASS benchmark; Graph Retrieval Augmented Generation using ground truth causal graphs.

Result: Causal reasoning remains broadly stable under quantization (NF4 <1% overall degradation); interventional queries (rung 2) most sensitive; counterfactual reasoning (rung 3) stable but shows heterogeneous weaknesses; CRASS benchmark shows near identical performance across precisions; graph augmentation improves NF4 interventional accuracy by +1.7%.

Conclusion: Causal reasoning is unexpectedly robust to 4-bit quantization; graph-structured augmentation can selectively reinforce interventional reasoning; current counterfactual benchmarks fail to capture deeper causal brittleness; provides empirical guidance for deploying efficient causal AI systems.

Abstract: Causal reasoning in Large Language Models spanning association, intervention, and counterfactual inference is essential for reliable decision making in high stakes settings. As deployment shifts toward edge and resource constrained environments, quantized models such as INT8 and NF4 are becoming standard. Yet the impact of precision reduction on formal causal reasoning is poorly understood. To our knowledge, this is the first study to systematically evaluate quantization effects across all three levels of Pearls Causal Ladder. Using a 3000 sample stratified CLadder benchmark, we find that rung level accuracy in Llama 3 8B remains broadly stable under quantization, with NF4 showing less than one percent overall degradation. Interventional queries at rung 2 are the most sensitive to precision loss, whereas counterfactual reasoning at rung 3 is comparatively stable but exhibits heterogeneous weaknesses across query types such as collider bias and backdoor adjustment. Experiments on the CRASS benchmark show near identical performance across precisions, indicating that existing commonsense counterfactual datasets lack the structural sensitivity needed to reveal quantization induced reasoning drift. We further evaluate Graph Retrieval Augmented Generation using ground truth causal graphs and observe a consistent improvement in NF4 interventional accuracy of plus 1.7 percent, partially offsetting compression related degradation. These results suggest that causal reasoning is unexpectedly robust to four bit quantization, graph structured augmentation can selectively reinforce interventional reasoning, and current counterfactual benchmarks fail to capture deeper causal brittleness. This work provides an initial empirical map of compressed causal reasoning and practical guidance for deploying efficient and structurally supported causal AI systems.

Method: The authors first analyze UT variants to identify key performance drivers, finding that recurrent inductive bias and strong nonlinear components are crucial. Based on this, they propose Universal Reasoning Model (URM) which enhances UT with short convolution and truncated backpropagation.

Result: URM achieves state-of-the-art performance: 53.8% pass@1 on ARC-AGI 1 and 16.0% pass@1 on ARC-AGI 2, substantially improving reasoning capabilities.

Conclusion: Performance gains in Universal Transformers for reasoning tasks primarily come from recurrent inductive bias and nonlinear components, not elaborate architectural designs. URM effectively leverages these insights to achieve SOTA results on ARC-AGI benchmarks.

Abstract: Universal transformers (UTs) have been widely used for complex reasoning tasks such as ARC-AGI and Sudoku, yet the specific sources of their performance gains remain underexplored. In this work, we systematically analyze UTs variants and show that improvements on ARC-AGI primarily arise from the recurrent inductive bias and strong nonlinear components of Transformer, rather than from elaborate architectural designs. Motivated by this finding, we propose the Universal Reasoning Model (URM), which enhances the UT with short convolution and truncated backpropagation. Our approach substantially improves reasoning performance, achieving state-of-the-art 53.8% pass@1 on ARC-AGI 1 and 16.0% pass@1 on ARC-AGI 2. Our code is avaliable at https://github.com/UbiquantAI/URM.

Method: 1) Fine-tunes LLaMA 3.2 3B using instruction-based learning on SentiFin dataset; 2) Implements RAG pipeline with dynamic multi-source contextual information selection via cosine similarity; 3) Adds feedback-driven module adjusting source reliability by comparing predicted sentiment with next-day stock returns; 4) Incorporates PPO reinforcement learning agent to generalize adaptive mechanism across temporal data.

Result: Experimental results on NIFTY 50 news headlines (2024-2025) show significant improvements in classification accuracy, F1-score, and market alignment over baseline models and static retrieval methods.

Conclusion: The framework successfully demonstrates the potential of combining instruction-tuned LLMs with dynamic feedback and reinforcement learning for robust, market-aware financial sentiment modeling that adapts to real-world market behavior.

Abstract: Financial sentiment analysis plays a crucial role in informing investment decisions, assessing market risk, and predicting stock price trends. Existing works in financial sentiment analysis have not considered the impact of stock prices or market feedback on sentiment analysis. In this paper, we propose an adaptive framework that integrates large language models (LLMs) with real-world stock market feedback to improve sentiment classification in the context of the Indian stock market. The proposed methodology fine-tunes the LLaMA 3.2 3B model using instruction-based learning on the SentiFin dataset. To enhance sentiment predictions, a retrieval-augmented generation (RAG) pipeline is employed that dynamically selects multi-source contextual information based on the cosine similarity of the sentence embeddings. Furthermore, a feedback-driven module is introduced that adjusts the reliability of the source by comparing predicted sentiment with actual next-day stock returns, allowing the system to iteratively adapt to market behavior. To generalize this adaptive mechanism across temporal data, a reinforcement learning agent trained using proximal policy optimization (PPO) is incorporated. The PPO agent learns to optimize source weighting policies based on cumulative reward signals from sentiment-return alignment. Experimental results on NIFTY 50 news headlines collected from 2024 to 2025 demonstrate that the proposed system significantly improves classification accuracy, F1-score, and market alignment over baseline models and static retrieval methods. The results validate the potential of combining instruction-tuned LLMs with dynamic feedback and reinforcement learning for robust, market-aware financial sentiment modeling.

Method: Two-stage training paradigm: first builds editing capability, then optimizes properties while reusing learned editing behaviors. Uses LLM agents that interact in multiple turns, invoking chemical tools for validity checking, property assessment, and similarity control.

Result: MolEditAgent-7B achieves 100, 95, and 98 valid add/delete/substitute edits, outperforming DeepSeek-R1. MolOptAgent-7B surpasses Claude 3.7 on LogP optimization and remains competitive on solubility while maintaining balanced performance across objectives.

Conclusion: Treating molecular design as a multi-step, tool-augmented process enables reliable and interpretable improvements, with MolAct being the first to formalize molecular design as an Agentic Reinforcement Learning problem.

Abstract: Molecular editing and optimization are multi-step problems that require iteratively improving properties while keeping molecules chemically valid and structurally similar. We frame both tasks as sequential, tool-guided decisions and introduce MolAct, an agentic reinforcement learning framework that employs a two-stage training paradigm: first building editing capability, then optimizing properties while reusing the learned editing behaviors. To the best of our knowledge, this is the first work to formalize molecular design as an Agentic Reinforcement Learning problem, where an LLM agent learns to interleave reasoning, tool-use, and molecular optimization. The framework enables agents to interact in multiple turns, invoking chemical tools for validity checking, property assessment, and similarity control, and leverages their feedback to refine subsequent edits. We instantiate the MolAct framework to train two model families: MolEditAgent for molecular editing tasks and MolOptAgent for molecular optimization tasks. In molecular editing, MolEditAgent-7B delivers 100, 95, and 98 valid add, delete, and substitute edits, outperforming strong closed “thinking” baselines such as DeepSeek-R1; MolEditAgent-3B approaches the performance of much larger open “thinking” models like Qwen3-32B-think. In molecular optimization, MolOptAgent-7B (trained on MolEditAgent-7B) surpasses the best closed “thinking” baseline (e.g., Claude 3.7) on LogP and remains competitive on solubility, while maintaining balanced performance across other objectives. These results highlight that treating molecular design as a multi-step, tool-augmented process is key to reliable and interpretable improvements.

Method: SACodec uses an asymmetric dual-quantizer with Semantic Anchoring mechanism. It employs a lightweight projector to align acoustic features with a frozen mHuBERT codebook for semantic quantization, then uses a residual activation module with SimVQ for acoustic detail recovery.

Result: At 1.5 kbps, SACodec establishes new state-of-the-art performance, with subjective listening tests showing reconstruction quality perceptually comparable to ground-truth audio, and tokens demonstrating substantially improved semantic richness in downstream tasks.

Conclusion: SACodec successfully addresses the fidelity-semantics trade-off in low-bitrate neural speech codecs through semantic anchoring, achieving both high perceptual quality and semantic richness at extremely low bitrates.

Abstract: Neural Speech Codecs face a fundamental trade-off at low bitrates: preserving acoustic fidelity often compromises semantic richness. To address this, we introduce SACodec, a novel codec built upon an asymmetric dual-quantizer that employs our proposed Semantic Anchoring mechanism. This design strategically decouples the quantization of Semantic and Acoustic details. The semantic anchoring is achieved via a lightweight projector that aligns acoustic features with a frozen, large-scale mHuBERT codebook, injecting linguistic priors while guaranteeing full codebook utilization. Sequentially, for acoustic details, a residual activation module with SimVQ enables a single-layer quantizer (acoustic path) to faithfully recover fine-grained information. At just 1.5 kbps, SACodec establishes a new state of the art by excelling in both fidelity and semantics: subjective listening tests confirm that its reconstruction quality is perceptually highly comparable to ground-truth audio, while its tokens demonstrate substantially improved semantic richness in downstream tasks.

Method: Two-step approach: music simplification followed by harmonization. Two solutions implemented using existing MidiBERT framework. Semi-supervised learning to reduce labeling effort while utilizing abundant classical music data.

Result: Solutions output practical and realistic samples with accurate reduction requiring only small post-processing adjustments. Forms groundwork for semi-supervised learning in automatic piano reduction.

Conclusion: Semi-supervised learning is effective for automatic piano reduction, producing practical results with minimal labeling. Provides foundation for future research to build more state-of-the-art solutions.

Abstract: In this study, we present a novel automatic piano reduction method with semi-supervised machine learning. Piano reduction is an important music transformation process, which helps musicians and composers as a musical sketch for performances and analysis. The automation of such is a highly challenging research problem but could bring huge conveniences as manually doing a piano reduction takes a lot of time and effort. While supervised machine learning is often a useful tool for learning input-output mappings, it is difficult to obtain a large quantity of labelled data. We aim to solve this problem by utilizing semi-supervised learning, so that the abundant available data in classical music can be leveraged to perform the task with little or no labelling effort. In this regard, we formulate a two-step approach of music simplification followed by harmonization. We further propose and implement two possible solutions making use of an existing machine learning framework – MidiBERT. We show that our solutions can output practical and realistic samples with an accurate reduction that needs only small adjustments in post-processing. Our study forms the groundwork for the use of semi-supervised learning in automatic piano reduction, where future researchers can take reference to produce more state-of-the-art results.

Method: Two-stage pipeline: 1) Create training data using LLM-generated lyrics paired with fixed melodies, 2) DiTSinger diffusion transformer with RoPE and qk-norm scaling, plus implicit alignment mechanism for phoneme-to-acoustic mapping without duration labels.

Result: Successfully synthesized over 500 hours of high-quality Chinese singing data and achieved scalable, alignment-free, high-fidelity SVS validated through extensive experiments.

Conclusion: The proposed approach enables scalable, alignment-free, and high-fidelity singing voice synthesis through data generation and architectural innovations.

Abstract: Recent progress in diffusion-based Singing Voice Synthesis (SVS) demonstrates strong expressiveness but remains limited by data scarcity and model scalability. We introduce a two-stage pipeline: a compact seed set of human-sung recordings is constructed by pairing fixed melodies with diverse LLM-generated lyrics, and melody-specific models are trained to synthesize over 500 hours of high-quality Chinese singing data. Building on this corpus, we propose DiTSinger, a Diffusion Transformer with RoPE and qk-norm, systematically scaled in depth, width, and resolution for enhanced fidelity. Furthermore, we design an implicit alignment mechanism that obviates phoneme-level duration labels by constraining phoneme-to-acoustic attention within character-level spans, thereby improving robustness under noisy or uncertain alignments. Extensive experiments validate that our approach enables scalable, alignment-free, and high-fidelity SVS.

Method: Created EnvSDD dataset with 45.25 hours of real and 316.7 hours of fake sound. Launched Environmental Sound Deepfake Detection Challenge with two tracks: ESDD in Unseen Generators and Black-Box Low-Resource ESDD.

Result: Proposed first large-scale curated dataset for environmental sound deepfake detection. The challenge will be held at ICASSP 2026 to address real-life detection challenges.

Conclusion: EnvSDD addresses the gap in environmental sound deepfake detection resources and the challenge promotes development of robust detection methods against potential misuse of audio generation technology.

Abstract: Recent advances in audio generation systems have enabled the creation of highly realistic and immersive soundscapes, which are increasingly used in film and virtual reality. However, these audio generators also raise concerns about potential misuse, such as generating deceptive audio content for fake videos and spreading misleading information. Existing datasets for environmental sound deepfake detection (ESDD) are limited in scale and audio types. To address this gap, we have proposed EnvSDD, the first large-scale curated dataset designed for ESDD, consisting of 45.25 hours of real and 316.7 hours of fake sound. Based on EnvSDD, we are launching the Environmental Sound Deepfake Detection Challenge. Specifically, we present two different tracks: ESDD in Unseen Generators and Black-Box Low-Resource ESDD, covering various challenges encountered in real-life scenarios. The challenge will be held in conjunction with the 2026 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2026).

Method: VCB Bench is built entirely on real human speech and evaluates LALMs from three perspectives: instruction following (including speech-level control), knowledge understanding (general knowledge, reasoning, daily dialogue), and robustness (stability under content, environment, and speaker perturbations).

Result: Experiments on representative LALMs reveal notable performance gaps and highlight future directions for improvement.

Conclusion: VCB Bench provides a reproducible and fine-grained evaluation framework with standardized methodology and practical insights for advancing Chinese voice conversational models.

Abstract: Recent advances in large audio language models (LALMs) have greatly enhanced multimodal conversational systems. However, existing benchmarks remain limited – they are mainly English-centric, rely on synthetic speech, and lack comprehensive, discriminative evaluation across multiple dimensions. To address these gaps, we present Voice Chat Bot Bench (VCB Bench) – a high-quality Chinese benchmark built entirely on real human speech. VCB Bench evaluates LALMs from three complementary perspectives: instruction following (including speech-level control beyond text commands), knowledge understanding (general knowledge, reasoning, and daily dialogue), and robustness (stability under perturbations in content, environment, and speaker traits). Experiments on representative LALMs reveal notable performance gaps and highlight future directions for improvement. VCB Bench provides a reproducible and fine-grained evaluation framework, offering standardized methodology and practical insights for advancing Chinese voice conversational models.

Method: Uses a modular ASR-LLM-SVS (Automatic Speech Recognition - Large Language Model - Singing Voice Synthesis) pipeline with configurable components including character personas, ASR/LLM backends, SVS models, melody sources, and voice profiles.

Result: Developed a plug-and-play web demo with modular, open-source code that supports customization and extension across different latency, quality, and musical style requirements.

Conclusion: SingingSDS enables singing-based dialogue responses for more engaging interactions in entertainment applications, with flexible configuration options and open-source availability for community use and extension.

Abstract: With recent advances in automatic speech recognition (ASR), large language models (LLMs), and text-to-speech (TTS) technologies, spoken dialogue systems (SDS) have become widely accessible. However, most existing SDS are limited to conventional spoken responses. We present SingingSDS, a cascaded SDS that responds through singing rather than speaking, fostering more affective, memorable, and pleasurable interactions in character-based roleplay and interactive entertainment scenarios. SingingSDS employs a modular ASR-LLM-SVS pipeline and supports a wide range of configurations across character personas, ASR and LLM backends, SVS models, melody sources, and voice profiles, tailored to different needs in terms of latency, quality, and musical style. SingingSDS is available as a plug-and-play web demo, featuring modular, open-source code that supports customization and extension. Demo: https://huggingface.co/spaces/espnet/SingingSDS. Code: https://github.com/SingingSDS/SingingSDS.

Method: Two-pronged approach: 1) Large-scale empirical study to characterize data scaling laws for SDD, quantifying source and generator diversity effects; 2) Proposed Diversity-Optimized Sampling Strategy (DOSS) with two implementations: DOSS-Select (pruning) and DOSS-Weight (re-weighting) for mixing heterogeneous data.

Result: DOSS-Select outperforms naive aggregation baseline using only 3% of total available data. Final model trained on 12k-hour curated data pool with optimal DOSS-Weight strategy achieves state-of-the-art performance, outperforming large-scale baselines with better data and model efficiency on public benchmarks and new commercial API challenge set.

Conclusion: Data-centric approaches, particularly through diversity-optimized sampling strategies, significantly improve speech deepfake detection generalization and efficiency, demonstrating that careful data composition is as important as model architecture for robust performance.

Abstract: Achieving robust generalization in speech deepfake detection (SDD) remains a primary challenge, as models often fail to detect unseen forgery methods. While research has focused on model-centric and algorithm-centric solutions, the impact of data composition is often underexplored. This paper proposes a data-centric approach, analyzing the SDD data landscape from two practical perspectives: constructing a single dataset and aggregating multiple datasets. To address the first perspective, we conduct a large-scale empirical study to characterize the data scaling laws for SDD, quantifying the impact of source and generator diversity. To address the second, we propose the Diversity-Optimized Sampling Strategy (DOSS), a principled framework for mixing heterogeneous data with two implementations: DOSS-Select (pruning) and DOSS-Weight (re-weighting). Our experiments show that DOSS-Select outperforms the naive aggregation baseline while using only 3% of the total available data. Furthermore, our final model, trained on a 12k-hour curated data pool using the optimal DOSS-Weight strategy, achieves state-of-the-art performance, outperforming large-scale baselines with greater data and model efficiency on both public benchmarks and a new challenge set of various commercial APIs.

Method: Hybrid feature extraction using MFCC and Wavelet Packet Transform (WPT), with GMM for speaker identification and HMM for speaker verification, tested on Voxforge and CSTR US KED Timit databases with noise evaluation.

Result: Experimental results show better performance for both speaker identification and verification tasks, with demonstrated noise robustness across different SNR levels.

Conclusion: The proposed wavelet packet-based hybrid feature extraction approach effectively improves text-independent speaker recognition performance and demonstrates good noise robustness.

Abstract: This paper proposes a novel Wavelet Packet based feature extraction approach for the task of text independent speaker recognition. The features are extracted by using the combination of Mel Frequency Cepstral Coefficient (MFCC) and Wavelet Packet Transform (WPT).Hybrid Features technique uses the advantage of human ear simulation offered by MFCC combining it with multi-resolution property and noise robustness of WPT. To check the validity of the proposed approach for the text independent speaker identification and verification we have used the Gaussian Mixture Model (GMM) and Hidden Markov Model (HMM) respectively as the classifiers. The proposed paradigm is tested on voxforge speech corpus and CSTR US KED Timit database. The paradigm is also evaluated after adding standard noise signal at different level of SNRs for evaluating the noise robustness. Experimental results show that better results are achieved for the tasks of both speaker identification as well as speaker verification.

Method: Proposes a parameter-efficient alternative to Neural CDEs that requires fewer parameters while maintaining the “Continuous RNN” analogy that Neural CDEs aspire to achieve.

Result: The proposed method achieves comparable performance to Neural CDEs but with significantly fewer parameters, making it more efficient for temporal sequence analysis tasks.

Conclusion: The parameter-efficient “Continuous RNN” alternative successfully addresses the main drawback of Neural CDEs while preserving their logical analogy and effectiveness for temporal sequence analysis.

Abstract: Neural Controlled Differential Equations (Neural CDEs, NCDEs) are a unique branch of methods, specifically tailored for analysing temporal sequences. However, they come with drawbacks, the main one being the number of parameters, required for the method’s operation. In this paper, we propose an alternative, parameter-efficient look at Neural CDEs. It requires much fewer parameters, while also presenting a very logical analogy as the “Continuous RNN”, which the Neural CDEs aspire to.

Method: Dual-loop architecture: behavior loop adjusts action preferences based on environmental rewards, while language loop updates external latent vectors by reflecting on semantic embeddings of generated text. This liberates latent vectors from static semantic representations.

Result: Agents’ latent spaces show clear convergence trajectories under reflection-driven updates with structured shifts at critical moments. The system demonstrates emergent ability to implicitly infer and adapt to emotional agents without shared rewards.

Conclusion: External latent spaces can provide language agents with low-cost, scalable, and interpretable abstract strategic representation without modifying model parameters, enabling continual strategy evolution.

Abstract: This study proposes a multi-agent language framework that enables continual strategy evolution without fine-tuning the language model’s parameters. The core idea is to liberate the latent vectors of abstract concepts from traditional static semantic representations, allowing them to be continuously updated through environmental interaction and reinforcement feedback. We construct a dual-loop architecture: the behavior loop adjusts action preferences based on environmental rewards, while the language loop updates the external latent vectors by reflecting on the semantic embeddings of generated text. Together, these mechanisms allow agents to develop stable and disentangled strategic styles over long-horizon multi-round interactions. Experiments show that agents’ latent spaces exhibit clear convergence trajectories under reflection-driven updates, along with structured shifts at critical moments. Moreover, the system demonstrates an emergent ability to implicitly infer and continually adapt to emotional agents, even without shared rewards. These results indicate that, without modifying model parameters, an external latent space can provide language agents with a low-cost, scalable, and interpretable form of abstract strategic representation.

Method: Comprehensive zero-training analysis using three transformer architectures on 12,279 authentic social media posts from major events, with statistical validation and introduction of four novel drift metrics compared against embedding-based baselines.

Result: Significant model instability with accuracy drops up to 23.4% during event-driven periods, maximum confidence drops of 13.0% (Bootstrap 95% CI: [9.1%, 16.5%]), and novel metrics outperforming baselines while maintaining computational efficiency for production.

Conclusion: Zero-training methodology enables immediate deployment for real-time sentiment monitoring, provides new insights into transformer behavior during dynamic content, and offers practical drift detection exceeding industry monitoring thresholds.

Abstract: We present a comprehensive zero-training temporal drift analysis of transformer-based sentiment models validated on authentic social media data from major real-world events. Through systematic evaluation across three transformer architectures and rigorous statistical validation on 12,279 authentic social media posts, we demonstrate significant model instability with accuracy drops reaching 23.4% during event-driven periods. Our analysis reveals maximum confidence drops of 13.0% (Bootstrap 95% CI: [9.1%, 16.5%]) with strong correlation to actual performance degradation. We introduce four novel drift metrics that outperform embedding-based baselines while maintaining computational efficiency suitable for production deployment. Statistical validation across multiple events confirms robust detection capabilities with practical significance exceeding industry monitoring thresholds. This zero-training methodology enables immediate deployment for real-time sentiment monitoring systems and provides new insights into transformer model behavior during dynamic content periods.

Method: Introduces a framework using LLMs as Goal-oriented Knowledge Curators (GKC) to convert laboratory, genomic, and medication data into high-fidelity, task-aligned features. Unlike generic embeddings, GKC produces objective-tailored representations and operates as an offline preprocessing step compatible with hospital informatics pipelines.

Result: Tested on a lung cancer cohort (N=184), GKC achieved mean AUROC of 0.803 (95% CI: 0.799-0.807), outperforming expert-engineered features, direct text embeddings, and end-to-end transformers. Ablation study confirmed complementary value of combining all three modalities.

Conclusion: Semantic representation quality is key for predictive accuracy in sparse clinical data. Reframing LLMs as knowledge curation engines rather than black-box predictors provides a scalable, interpretable, workflow-compatible pathway for AI-driven oncology decision support.

Abstract: Accurate prediction of treatment outcomes in lung cancer remains challenging due to the sparsity, heterogeneity, and contextual overload of real-world electronic health data. Traditional models often fail to capture semantic information across multimodal streams, while large-scale fine-tuning approaches are impractical in clinical workflows. We introduce a framework that uses Large Language Models (LLMs) as Goal-oriented Knowledge Curators (GKC) to convert laboratory, genomic, and medication data into high-fidelity, task-aligned features. Unlike generic embeddings, GKC produces representations tailored to the prediction objective and operates as an offline preprocessing step that integrates naturally into hospital informatics pipelines. Using a lung cancer cohort (N=184), we benchmarked GKC against expert-engineered features, direct text embeddings, and an end-to-end transformer. Our approach achieved a mean AUROC of 0.803 (95% CI: 0.799-0.807) and outperformed all baselines. An ablation study further confirmed the complementary value of combining all three modalities. These results show that the quality of semantic representation is a key determinant of predictive accuracy in sparse clinical data settings. By reframing LLMs as knowledge curation engines rather than black-box predictors, this work demonstrates a scalable, interpretable, and workflow-compatible pathway for advancing AI-driven decision support in oncology.

Method: Proposes a comprehensive framework with: (1) quantitative metrics (0-1 scale) to measure alignment depth across token positions; (2) real-time detection methods for identifying shallow alignment during training; (3) specialized analysis tools for visualization and recovery prediction; and (4) adaptive mitigation strategies that automatically distinguish forgetting types and promote deep alignment.

Result: Experiments on multiple datasets and model architectures (Qwen2.5-3B to Qwen2.5-32B) demonstrate 86.2-90.6% identification accuracy for shallow alignment, and promoting deep alignment improves robustness against forgetting by 3.3-7.1% over baselines.

Conclusion: The shallow vs deep alignment framework provides the first quantitative characterization of alignment depth, explains why spurious forgetting occurs and is reversible, and offers practical tools to measure, detect, and mitigate catastrophic forgetting in continual learning for LLMs.

Abstract: Catastrophic forgetting remains a fundamental challenge in continual learning for large language models. Recent work revealed that performance degradation may stem from spurious forgetting caused by task alignment disruption rather than true knowledge loss. However, this work only qualitatively describes alignment, relies on post-hoc analysis, and lacks automatic distinction mechanisms. We introduce the shallow versus deep alignment framework, providing the first quantitative characterization of alignment depth. We identify that current task alignment approaches suffer from shallow alignment - maintained only over the first few output tokens (approximately 3-5) - making models vulnerable to forgetting. This explains why spurious forgetting occurs, why it is reversible, and why fine-tuning attacks are effective. We propose a comprehensive framework addressing all gaps: (1) quantitative metrics (0-1 scale) to measure alignment depth across token positions; (2) real-time detection methods for identifying shallow alignment during training; (3) specialized analysis tools for visualization and recovery prediction; and (4) adaptive mitigation strategies that automatically distinguish forgetting types and promote deep alignment. Extensive experiments on multiple datasets and model architectures (Qwen2.5-3B to Qwen2.5-32B) demonstrate 86.2-90.6% identification accuracy and show that promoting deep alignment improves robustness against forgetting by 3.3-7.1% over baselines.

Method: SHRP introduces Expert Attention where each attention head is treated as an independent expert, followed by a shared expander feed-forward network. It uses a unified Top-1 usage-driven mechanism for joint dynamic routing during training and deterministic pruning at deployment.

Result: On GLUE benchmark with BERT-base, SHRP achieves 93% original accuracy with 48% parameter reduction. In extreme compression (11/12 layers pruned), maintains 84% accuracy with 4.2x throughput gain and reduces computation to 11.5% of original FLOPs.

Conclusion: SHRP effectively compresses Transformer encoders by pruning redundant attention heads while preserving accuracy, demonstrating practical utility for large-scale, latency-sensitive web deployments through significant parameter and computation reduction.

Abstract: Transformer encoders are widely deployed in large-scale web services for natural language understanding tasks such as text classification, semantic retrieval, and content ranking. However, their high inference latency and memory consumption pose significant challenges for real-time serving and scalability. These limitations stem largely from architectural redundancy, particularly in the attention module. The inherent parameter redundancy of the attention mechanism, coupled with the fact that its attention heads operate with a degree of independence, makes it particularly amenable to structured model compression. In this paper, we propose SHRP (Specialized Head Routing and Pruning), a novel structured pruning framework that automatically identifies and removes redundant attention heads while preserving most of the model’s accuracy and compatibility. SHRP introduces Expert Attention, a modular design that treats each attention head as an independent expert, followed by a lightweight shared expander feed-forward network that refines their outputs. The framework employs a unified Top-1 usage-driven mechanism to jointly perform dynamic routing during training and deterministic pruning at deployment. Experimental results on the GLUE benchmark using a BERT-base encoder show that SHRP achieves 93% of the original model accuracy while reducing parameters by 48 percent. Under an extreme compression scenario where 11/12 of the layers are pruned, the model still maintains 84% accuracy and delivers a 4.2x throughput gain while reducing computation to as low as 11.5 percent of the original FLOPs, demonstrating its practical utility for large-scale and latency-sensitive web deployments.

Method: Gate-Norm uses a one-shot, weight-only criterion that ranks attention sublayers by query-key coupling and removes the least coupled ones. It requires no calibration data, no forward passes, no fine-tuning, and no specialized kernels.

Result: On 40-layer, 13B-parameter LLaMA models, Gate-Norm prunes 8-16 attention sublayers in under a second, achieving up to 1.30x higher inference throughput while keeping average zero-shot accuracy within 2% of baseline across multiple benchmarks.

Conclusion: Gate-Norm matches data-driven pruning methods in accuracy while being ~1000x faster to score layers, enabling practical, data-free compression of LLMs through efficient attention layer pruning.

Abstract: Many self-attention sublayers in large language models (LLMs) can be removed with little to no loss. We attribute this to the Attention Suppression Hypothesis: during pre-training, some deep attention layers learn to mute their own contribution, leaving the residual stream and the MLP to carry the representation. We propose Gate-Norm, a one-shot, weight-only criterion that ranks attention sublayers by query–key coupling and removes the least coupled ones, requiring no calibration data, no forward passes, no fine-tuning, and no specialized kernels. On 40-layer, 13B-parameter LLaMA models, Gate-Norm prunes the model in under a second. Pruning $8$–$16$ attention sublayers yields up to $1.30\times$ higher inference throughput while keeping average zero-shot accuracy within $2%$ of the unpruned baseline across BoolQ, RTE, HellaSwag, WinoGrande, ARC-Easy/Challenge, and OpenBookQA. Across these settings, Gate-Norm matches data-driven pruning methods in accuracy while being $\sim 1000\times$ faster to score layers, enabling practical, data-free compression of LLMs.

Method: Uses Scientific ML frameworks in Julia: Neural ODEs and Universal Differential Equations (UDEs) to predict n-body system dynamics. Employs synthetically created noisy data to simulate real-world observational limitations. Determines forecasting breakdown point - minimum training data needed for accurate predictions.

Result: UDE model is much more data efficient, requiring only 20% of data for correct forecasting, while Neural ODE requires 90% of data. UDEs demonstrate superior performance in handling noisy observational data with limited training.

Conclusion: Scientific ML approaches, particularly Universal Differential Equations, offer significant advantages for n-body problem forecasting by embedding physical laws, improving interpretability, and dramatically reducing data requirements compared to traditional ML methods and even Neural ODEs.

Abstract: The n body problem, fundamental to astrophysics, simulates the motion of n bodies acting under the effect of their own mutual gravitational interactions. Traditional machine learning models that are used for predicting and forecasting trajectories are often data intensive black box models, which ignore the physical laws, thereby lacking interpretability. Whereas Scientific Machine Learning ( Scientific ML ) directly embeds the known physical laws into the machine learning framework. Through robust modelling in the Julia programming language, our method uses the Scientific ML frameworks: Neural ordinary differential equations (NODEs) and Universal differential equations (UDEs) to predict and forecast the system dynamics. In addition, an essential component of our analysis involves determining the forecasting breakdown point, which is the smallest possible amount of training data our models need to predict future, unseen data accurately. We employ synthetically created noisy data to simulate real-world observational limitations. Our findings indicate that the UDE model is much more data efficient, needing only 20% of data for a correct forecast, whereas the Neural ODE requires 90%.

Method: Proposed Q-RUN (quantum-inspired data re-uploading network), which mathematically adapts the DRQC paradigm to classical neural networks. It serves as a drop-in replacement for standard fully connected layers while retaining the Fourier-expressive properties of quantum models.

Result: Q-RUN outperforms both standard fully connected layers and state-of-the-art neural network layers, reducing model parameters while decreasing error by approximately 1-3 orders of magnitude on certain tasks. It improves performance across various neural architectures.

Conclusion: This work demonstrates that quantum machine learning principles can guide the design of more expressive classical AI models, showing how quantum-inspired approaches can enhance neural network performance without requiring quantum hardware.

Abstract: Data re-uploading quantum circuits (DRQC) are a key approach to implementing quantum neural networks and have been shown to outperform classical neural networks in fitting high-frequency functions. However, their practical application is limited by the scalability of current quantum hardware. In this paper, we introduce the mathematical paradigm of DRQC into classical models by proposing a quantum-inspired data re-uploading network (Q-RUN), which retains the Fourier-expressive advantages of quantum models without any quantum hardware. Experimental results demonstrate that Q-RUN delivers superior performance across both data modeling and predictive modeling tasks. Compared to the fully connected layers and the state-of-the-art neural network layers, Q-RUN reduces model parameters while decreasing error by approximately one to three orders of magnitude on certain tasks. Notably, Q-RUN can serve as a drop-in replacement for standard fully connected layers, improving the performance of a wide range of neural architectures. This work illustrates how principles from quantum machine learning can guide the design of more expressive artificial intelligence.

Method: Created MaskOpt dataset from real IC designs at 45nm node with 104,714 metal-layer tiles and 121,952 via-layer tiles. Tiles are clipped at standard-cell placements to preserve cell information and exploit repeated logic gates. Supports different context window sizes to capture optical proximity effects.

Result: Evaluated state-of-the-art deep learning models, revealing distinct trade-offs across baseline models. Context size analysis and input ablation studies confirmed the importance of surrounding geometries and cell-aware inputs for accurate mask generation.

Conclusion: MaskOpt advances deep learning for practical mask optimization by providing a real-world benchmark that captures cell hierarchy and context awareness, addressing limitations of existing synthetic datasets.

Abstract: As integrated circuit (IC) dimensions shrink below the lithographic wavelength, optical lithography faces growing challenges from diffraction and process variability. Model-based optical proximity correction (OPC) and inverse lithography technique (ILT) remain indispensable but computationally expensive, requiring repeated simulations that limit scalability. Although deep learning has been applied to mask optimization, existing datasets often rely on synthetic layouts, disregard standard-cell hierarchy, and neglect the surrounding contexts around the mask optimization targets, thereby constraining their applicability to practical mask optimization. To advance deep learning for cell- and context-aware mask optimization, we present MaskOpt, a large-scale benchmark dataset constructed from real IC designs at the 45$\mathrm{nm}$ node. MaskOpt includes 104,714 metal-layer tiles and 121,952 via-layer tiles. Each tile is clipped at a standard-cell placement to preserve cell information, exploiting repeated logic gate occurrences. Different context window sizes are supported in MaskOpt to capture the influence of neighboring shapes from optical proximity effects. We evaluate state-of-the-art deep learning models for IC mask optimization to build up benchmarks, and the evaluation results expose distinct trade-offs across baseline models. Further context size analysis and input ablation studies confirm the importance of both surrounding geometries and cell-aware inputs in achieving accurate mask generation.

Method: Proposes a Dual-State Architecture separating workflow state (deterministic control flow) from environment state (stochastic generation). Uses Atomic Action Pairs that couple generation with verification as indivisible transactions, where Guard Functions act as sensing actions that project probabilistic outputs onto observable workflow state.

Result: Validated on three code generation tasks across 13 LLMs (1.3B-15B parameters). For qualified instruction-following models, task success rates improved by up to 66 percentage points at 1.2-2.1× baseline computational cost.

Conclusion: Architectural constraints can substitute for parameter scale in achieving reliable code generation. Properly separating deterministic control from stochastic LLM generation significantly improves reliability without requiring larger models.

Abstract: Current approaches to AI coding agents appear to blur the lines between the Large Language Model (LLM) and the agent itself, asking the LLM to make decisions best left to deterministic processes. This leads to systems prone to stochastic failures such as gaming unit tests or hallucinating syntax. Drawing on established software engineering practices that provide deterministic frameworks for managing unpredictable processes, this paper proposes setting the control boundary such that the LLM is treated as a component of the environment environment – preserving its creative stochasticity – rather than the decision-making agent. A \textbf{Dual-State Architecture} is formalized, separating workflow state (deterministic control flow) from environment state (stochastic generation). \textbf{Atomic Action Pairs} couple generation with verification as indivisible transactions, where \textbf{Guard Functions} act as sensing actions that project probabilistic outputs onto observable workflow state. The framework is validated on three code generation tasks across 13 LLMs (1.3B–15B parameters). For qualified instruction-following models, task success rates improved by up to 66 percentage points at 1.2–2.1$\times$ baseline computational cost. The results suggest that architectural constraints can substitute for parameter scale in achieving reliable code generation.

Method: The authors introduce DominanceBench to systematically analyze DvD imbalance. They examine causes from both data and architectural perspectives, conduct experiments on training data diversity, analyze cross-attention dynamics across diffusion timesteps, and perform head ablation studies to understand distributed attention mechanisms.

Result: Limited instance diversity in training data exacerbates inter-concept interference. Dominant tokens rapidly saturate attention and progressively suppress others across diffusion timesteps. The DvD behavior arises from distributed attention mechanisms across multiple heads rather than isolated components.

Conclusion: The findings provide key insights into generative collapse in diffusion models and advance toward more reliable and controllable text-to-image generation by understanding and addressing the DvD imbalance phenomenon.

Abstract: Text-to-image diffusion models have drawn significant attention for their ability to generate diverse and high-fidelity images. However, when generating from multi-concept prompts, one concept token often dominates the generation, suppressing the others-a phenomenon we term the Dominant-vs-Dominated (DvD) imbalance. To systematically analyze this imbalance, we introduce DominanceBench and examine its causes from both data and architectural perspectives. Through various experiments, we show that the limited instance diversity in training data exacerbates the inter-concept interference. Analysis of cross-attention dynamics further reveals that dominant tokens rapidly saturate attention, progressively suppressing others across diffusion timesteps. In addition, head ablation studies show that the DvD behavior arises from distributed attention mechanisms across multiple heads. Our findings provide key insights into generative collapse, advancing toward more reliable and controllable text-to-image generation.

Method: The authors first analyze theoretical limitations of existing forward-only approaches, then design a forward-only algorithm equivalent to backpropagation under linear and orthogonal assumptions. By relaxing the linear assumption, they develop FOTON (Forward-Only Training of Orthogonal Networks) that bridges the gap with backpropagation.

Result: FOTON outperforms PEPITA and enables training of neural networks of any depth without backward passes. It also shows promising performance on convolutional networks, opening avenues for application to more advanced architectures.

Conclusion: FOTON provides a viable alternative to backpropagation that eliminates the computational burden of backward passes while maintaining the ability to train deep networks effectively, with code made publicly available.

Abstract: Backpropagation is still the de facto algorithm used today to train neural networks. With the exponential growth of recent architectures, the computational cost of this algorithm also becomes a burden. The recent PEPITA and forward-only frameworks have proposed promising alternatives, but they failed to scale up to a handful of hidden layers, yet limiting their use. In this paper, we first analyze theoretically the main limitations of these approaches. It allows us the design of a forward-only algorithm, which is equivalent to backpropagation under the linear and orthogonal assumptions. By relaxing the linear assumption, we then introduce FOTON (Forward-Only Training of Orthogonal Networks) that bridges the gap with the backpropagation algorithm. Experimental results show that it outperforms PEPITA, enabling us to train neural networks of any depth, without the need for a backward pass. Moreover its performance on convolutional networks clearly opens up avenues for its application to more advanced architectures. The code is open-sourced at https://github.com/p0lcAi/FOTON .

Method: Uses a Conditional Variational Autoencoder (CVAE) architecture for synthesizing realistic clinical records that are coherent with real-world observations. The approach aims to generate synthetic data that maintains the statistical properties and relationships of real clinical data.

Result: The proposed CVAE architecture successfully generates coherent and realistic synthetic data. Using this synthetic data improves the accuracy of various classifiers for cardiac risk detection, outperforming state-of-the-art deep learning approaches for synthetic data generation.

Conclusion: The CVAE-based synthetic data generation approach effectively addresses limitations of real-world medical databases by increasing dataset size and diversity, reducing the need for potentially hazardous diagnostic procedures like exercise stress testing, and improving cardiac risk prediction model performance.

Abstract: Cardiac rehabilitation constitutes a structured clinical process involving multiple interdependent phases, individualized medical decisions, and the coordinated participation of diverse healthcare professionals. This sequential and adaptive nature enables the program to be modeled as a business process, thereby facilitating its analysis. Nevertheless, studies in this context face significant limitations inherent to real-world medical databases: data are often scarce due to both economic costs and the time required for collection; many existing records are not suitable for specific analytical purposes; and, finally, there is a high prevalence of missing values, as not all patients undergo the same diagnostic tests. To address these limitations, this work proposes an architecture based on a Conditional Variational Autoencoder (CVAE) for the synthesis of realistic clinical records that are coherent with real-world observations. The primary objective is to increase the size and diversity of the available datasets in order to enhance the performance of cardiac risk prediction models and to reduce the need for potentially hazardous diagnostic procedures, such as exercise stress testing. The results demonstrate that the proposed architecture is capable of generating coherent and realistic synthetic data, whose use improves the accuracy of the various classifiers employed for cardiac risk detection, outperforming state-of-the-art deep learning approaches for synthetic data generation.

Method: Proposes Dynamic Conflict-Consensus Framework (DCCF) with three key components: 1) Decouples inputs into independent Fact and Sentiment spaces to separate objective mismatches from emotional dissonance; 2) Uses physics-inspired feature dynamics to iteratively polarize representations and actively extract maximally informative conflicts; 3) Employs conflict-consensus mechanism to standardize local discrepancies against global context for robust deliberative judgment.

Result: Extensive experiments on three real-world datasets show DCCF consistently outperforms state-of-the-art baselines, achieving an average accuracy improvement of 3.52%.

Conclusion: DCCF introduces a paradigm shift from consistency-seeking to inconsistency-seeking in multimodal fake news detection, demonstrating that actively amplifying contradictions rather than suppressing them leads to more effective detection of fabricated content.

Abstract: Prevalent multimodal fake news detection relies on consistency-based fusion, yet this paradigm fundamentally misinterprets critical cross-modal discrepancies as noise, leading to over-smoothing, which dilutes critical evidence of fabrication. Mainstream consistency-based fusion inherently minimizes feature discrepancies to align modalities, yet this approach fundamentally fails because it inadvertently smoothes out the subtle cross-modal contradictions that serve as the primary evidence of fabrication. To address this, we propose the Dynamic Conflict-Consensus Framework (DCCF), an inconsistency-seeking paradigm designed to amplify rather than suppress contradictions. First, DCCF decouples inputs into independent Fact and Sentiment spaces to distinguish objective mismatches from emotional dissonance. Second, we employ physics-inspired feature dynamics to iteratively polarize these representations, actively extracting maximally informative conflicts. Finally, a conflict-consensus mechanism standardizes these local discrepancies against the global context for robust deliberative judgment.Extensive experiments conducted on three real world datasets demonstrate that DCCF consistently outperforms state-of-the-art baselines, achieving an average accuracy improvement of 3.52%.

Method: HyDRA implements hierarchical optimization (coarse-grained rank assignment to different layers and fine-grained rank adjustment within individual layers) and dynamic adjustment (end-to-end automatic optimization using a lightweight performance model to determine and adjust ranks during fine-tuning).

Result: HyDRA consistently outperforms baselines, achieving 4.7% improvement across various model sizes without increasing trainable parameters. In some tasks, it even surpasses full-parameter fine-tuning.

Conclusion: HyDRA provides an effective parameter-efficient fine-tuning solution for mobile VLMs that addresses the limitations of standard LoRA through hierarchical and dynamic rank scheduling.

Abstract: Vision Language Models (VLMs) have undergone significant advancements, particularly with the emergence of mobile-oriented VLMs, which offer a wide range of application scenarios. However, the substantial computational requirements for training these models present a significant obstacle to their practical application. To address this issue, Low-Rank Adaptation (LoRA) has been proposed. Nevertheless, the standard LoRA with a fixed rank lacks sufficient capability for training mobile VLMs that process both text and image modalities. In this work, we introduce HyDRA, a parameter-efficient fine-tuning framework designed to implement hierarchical and dynamic rank scheduling for mobile VLMs. This framework incorporates two essential optimization strategies: (1) hierarchical optimization, which involves a coarse-grained approach that assigns different ranks to various layers, as well as a fine-grained method that adjusts ranks within individual layers, and (2) dynamic adjustment, which employs an end-to-end automatic optimization using a lightweight performance model to determine and adjust ranks during the fine-tuning process. Comprehensive experiments conducted on popular benchmarks demonstrate that HyDRA consistently outperforms the baseline, achieving a 4.7% improvement across various model sizes without increasing the number of trainable parameters. In some tasks, it even surpasses full-parameter fine-tuning.

Method: Created a unified framework to evaluate recent VLM-based reward models with identical backbones, finetuning data, and evaluation environments using Meta-World tasks. Assessed modeling accuracy through consistency with ground truth reward and correlation with expert progress.

Result: A simple triplet loss outperformed state-of-the-art methods, suggesting that much of the improvements in recent approaches could be attributed to differences in data and architectures rather than the learning objectives themselves.

Conclusion: The learning objective itself may be less critical than previously thought, and simpler approaches like triplet loss can be effective when evaluated under fair, controlled conditions.

Abstract: Learning generalizable reward functions is a core challenge in embodied intelligence. Recent work leverages contrastive vision language models (VLMs) to obtain dense, domain-agnostic rewards without human supervision. These methods adapt VLMs into reward models through increasingly complex learning objectives, yet meaningful comparison remains difficult due to differences in training data, architectures, and evaluation settings. In this work, we isolate the impact of the learning objective by evaluating recent VLM-based reward models under a unified framework with identical backbones, finetuning data, and evaluation environments. Using Meta-World tasks, we assess modeling accuracy by measuring consistency with ground truth reward and correlation with expert progress. Remarkably, we show that a simple triplet loss outperforms state-of-the-art methods, suggesting that much of the improvements in recent approaches could be attributed to differences in data and architectures.

Method: PHOTON maintains a hierarchy of latent streams: bottom-up encoder compresses tokens into low-rate contextual states, while lightweight top-down decoders reconstruct fine-grained token representations.

Result: PHOTON offers superior throughput-quality trade-off compared to Transformer-based language models, with significant advantages in long-context and multi-query tasks, reducing decode-time KV-cache traffic by up to 1000× higher throughput per unit memory.

Conclusion: The hierarchical approach of PHOTON effectively addresses the memory-bound limitations of Transformer inference, enabling more efficient long-context processing through vertical multi-resolution context access.

Abstract: Transformers operate as horizontal token-by-token scanners; at each generation step, the model attends to an ever-growing sequence of token-level states. This access pattern increases prefill latency and makes long-context decoding increasingly memory-bound, as KV-cache reads and writes dominate inference throughput rather than arithmetic computation. We propose Parallel Hierarchical Operation for Top-down Networks (PHOTON), a hierarchical autoregressive model that replaces flat scanning with vertical, multi-resolution context access. PHOTON maintains a hierarchy of latent streams: a bottom-up encoder progressively compresses tokens into low-rate contextual states, while lightweight top-down decoders reconstruct fine-grained token representations. Experimental results show that PHOTON is superior to competitive Transformer-based language models regarding the throughput-quality trade-off, offering significant advantages in long-context and multi-query tasks. This reduces decode-time KV-cache traffic, yielding up to $10^{3}\times$ higher throughput per unit memory.

Method: Created FEM-Bench 2025 with introductory but nontrivial computational mechanics tasks aligned with first graduate course material. Tasks capture essential numerical/physical modeling challenges while representing only a fraction of the discipline’s complexity.

Result: Best performing model (Gemini 3 Pro) completed 30/33 tasks at least once and 26/33 tasks all five times in function writing. GPT-5 had 73.8% Average Joint Success Rate in unit test writing. Other models showed broad performance variation.

Conclusion: FEM-Bench establishes a structured foundation for evaluating AI-generated scientific code. Current LLMs don’t reliably solve even introductory computational mechanics tasks, highlighting the need for continued progress tracking through increasingly sophisticated benchmark iterations.

Abstract: As LLMs advance their reasoning capabilities about the physical world, the absence of rigorous benchmarks for evaluating their ability to generate scientifically valid physical models has become a critical gap. Computational mechanics, which develops and applies mathematical models and numerical methods to predict the behavior of physical systems under forces, deformation, and constraints, provides an ideal foundation for structured scientific reasoning evaluation. Problems follow clear mathematical structure, enforce strict physical and numerical constraints, and support objective verification. The discipline requires constructing explicit models of physical systems and reasoning about geometry, spatial relationships, and material behavior, connecting directly to emerging AI goals in physical reasoning and world modeling. We introduce FEM-Bench, a computational mechanics benchmark designed to evaluate the ability of LLMs to generate correct finite element method (FEM) and related code. FEM-Bench 2025 contains a suite of introductory but nontrivial tasks aligned with material from a first graduate course on computational mechanics. These tasks capture essential numerical and physical modeling challenges while representing only a small fraction of the complexity present in the discipline. Despite their simplicity, state-of-the-art LLMs do not reliably solve all of them. In a five attempt run, the best performing model at function writing, Gemini 3 Pro, completed 30/33 tasks at least once and 26/33 tasks all five times. The best performing model at unit test writing, GPT-5, had an Average Joint Success Rate of 73.8%. Other popular models showed broad performance variation. FEM-Bench establishes a structured foundation for evaluating AI-generated scientific code, and future iterations will incorporate increasingly sophisticated tasks to track progress as models evolve.

Method: 1) Derive theoretical Lipschitz constants for aggregation methods in multimodal autoencoders; 2) Introduce regularized attention-based fusion method based on theoretical analysis; 3) Empirically validate findings by estimating Lipschitz constants across multiple trials and fusion strategies.

Result: Proposed fusion function aligns with theoretical predictions and outperforms existing strategies in consistency, convergence speed, and accuracy. Empirical validation confirms theoretical findings.

Conclusion: Provides theoretical foundation for understanding fusion in multimodal autoencoders and contributes a solution for enhancing their performance through improved stability and training characteristics.

Abstract: In recent years, the development of multimodal autoencoders has gained significant attention due to their potential to handle multimodal complex data types and improve model performance. Understanding the stability and robustness of these models is crucial for optimizing their training, architecture, and real-world applicability. This paper presents an analysis of Lipschitz properties in multimodal autoencoders, combining both theoretical insights and empirical validation to enhance the training stability of these models. We begin by deriving the theoretical Lipschitz constants for aggregation methods within the multimodal autoencoder framework. We then introduce a regularized attention-based fusion method, developed based on our theoretical analysis, which demonstrates improved stability and performance during training. Through a series of experiments, we empirically validate our theoretical findings by estimating the Lipschitz constants across multiple trials and fusion strategies. Our results demonstrate that our proposed fusion function not only aligns with theoretical predictions but also outperforms existing strategies in terms of consistency, convergence speed, and accuracy. This work provides a solid theoretical foundation for understanding fusion in multimodal autoencoders and contributes a solution for enhancing their performance.

Method: Define a robustness property for early exit architectures, use off-the-shelf solvers for verification, develop a baseline algorithm with early stopping strategy and heuristic optimizations that maintain soundness and completeness.

Result: Experiments on multiple benchmarks validate the framework’s effectiveness. Early exits not only accelerate inference but also enhance verifiability, enabling more queries to be solved in less time compared to standard networks.

Conclusion: Early exit architectures offer a beneficial trade-off between accuracy and efficiency, improving both inference speed and verifiability. The proposed verification framework helps users navigate this trade-off with robustness analysis.

Abstract: Ensuring the safety and efficiency of AI systems is a central goal of modern research. Formal verification provides guarantees of neural network robustness, while early exits improve inference efficiency by enabling intermediate predictions. Yet verifying networks with early exits introduces new challenges due to their conditional execution paths. In this work, we define a robustness property tailored to early exit architectures and show how off-the-shelf solvers can be used to assess it. We present a baseline algorithm, enhanced with an early stopping strategy and heuristic optimizations that maintain soundness and completeness. Experiments on multiple benchmarks validate our framework’s effectiveness and demonstrate the performance gains of the improved algorithm. Alongside the natural inference acceleration provided by early exits, we show that they also enhance verifiability, enabling more queries to be solved in less time compared to standard networks. Together with a robustness analysis, we show how these metrics can help users navigate the inherent trade-off between accuracy and efficiency.

Method: Construct datasets of causal graphs and queries across associational, interventional, and counterfactual levels with varying structural complexity. Fine-tune Qwen-2.5-Instruct models (3B-32B) using RLVR vs SFT, varying training query levels included.

Result: RLVR yields stronger within-level and across-level generalization than SFT, but only for specific model size/training query level combinations. RLVR’s effectiveness depends on initial reasoning competence - with sufficient competence, RLVR improves marginalization strategy and reduces intermediate probability calculation errors.

Conclusion: RLVR can improve specific causal reasoning subskills, but its benefits emerge only when the model has sufficient initial competence, showing RLVR’s effectiveness is conditional on model capabilities.

Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as a promising paradigm for post-training large language models (LLMs) on complex reasoning tasks. Yet, the conditions under which RLVR yields robust generalization remain poorly understood. This paper provides an empirical study of RLVR generalization in the setting of probabilistic inference over causal graphical models. This setting offers two natural axes along which to examine generalization: (i) the level of the probabilistic query – associational, interventional, or counterfactual – and (ii) the structural complexity of the query, measured by the size of its relevant subgraph. We construct datasets of causal graphs and queries spanning these difficulty axes and fine-tune Qwen-2.5-Instruct models using RLVR or supervised fine-tuning (SFT). We vary both the model scale (3B-32B) and the query level included in training. We find that RLVR yields stronger within-level and across-level generalization than SFT, but only for specific combinations of model size and training query level. Further analysis shows that RLVR’s effectiveness depends on the model’s initial reasoning competence. With sufficient initial competence, RLVR improves an LLM’s marginalization strategy and reduces errors in intermediate probability calculations, producing substantial accuracy gains, particularly on more complex queries. These findings show that RLVR can improve specific causal reasoning subskills, with its benefits emerging only when the model has sufficient initial competence.

Method: TS-Arena implements a pre-registration mechanism on live data streams, treating the genuinely unknown future as the definitive test environment. It establishes a moving temporal frontier with strict global temporal splits to prevent historical contamination.

Result: The platform provides a sustainable infrastructure for comparing foundation models under real-world constraints, initially applied within the energy sector, with a prototype available on Hugging Face.

Conclusion: TS-Arena restores operational integrity to forecasting evaluation by ensuring test targets remain physically non-existent during inference, enabling authentic assessment of model generalization capabilities.

Abstract: While Time Series Foundation Models (TSFMs) offer transformative capabilities for forecasting, they simultaneously risk triggering a fundamental evaluation crisis. This crisis is driven by information leakage due to overlapping training and test sets across different models, as well as the illegitimate transfer of global patterns to test data. While the ability to learn shared temporal dynamics represents a primary strength of these models, their evaluation on historical archives often permits the exploitation of observed global shocks, which violates the independence required for valid benchmarking. We introduce TS-Arena, a platform that restores the operational integrity of forecasting by treating the genuinely unknown future as the definitive test environment. By implementing a pre-registration mechanism on live data streams, the platform ensures that evaluation targets remain physically non-existent during inference, thereby enforcing a strict global temporal split. This methodology establishes a moving temporal frontier that prevents historical contamination and provides an authentic assessment of model generalization. Initially applied within the energy sector, TS-Arena provides a sustainable infrastructure for comparing foundation models under real-world constraints. A prototype of the platform is available at https://huggingface.co/spaces/DAG-UPB/TS-Arena.

Method: Introduced two technical innovations: Expected Prediction Entropy (EPE) for evaluating survival models predicting hazard functions, and Conditional Rank Statistics (CRS) for quantifying individual deviation from subgroup survival distributions. Developed eight algorithms for Cox subgroup discovery, with main algorithm combining EPE and CRS.

Result: Theoretical analysis shows EPE and CRS solve problems with existing metrics. Empirical evaluation on synthetic and real data demonstrates recovery of ground-truth subgroups in well-specified cases and better model fit than naive Cox models on whole datasets. NASA jet engine case study reveals known nonlinearities and validates practical design choices.

Conclusion: The paper presents the first comprehensive approach to subgroup discovery for survival analysis, with novel metrics and algorithms that effectively identify interpretable subgroups where Cox models perform well, validated through theory, experiments, and real-world application.

Abstract: We study the problem of subgroup discovery for survival analysis, where the goal is to find an interpretable subset of the data on which a Cox model is highly accurate. Our work is the first to study this particular subgroup problem, for which we make several contributions. Subgroup discovery methods generally require a “quality function” in order to sift through and select the most advantageous subgroups. We first examine why existing natural choices for quality functions are insufficient to solve the subgroup discovery problem for the Cox model. To address the shortcomings of existing metrics, we introduce two technical innovations: the expected prediction entropy (EPE), a novel metric for evaluating survival models which predict a hazard function; and the conditional rank statistics (CRS), a statistical object which quantifies the deviation of an individual point to the distribution of survival times in an existing subgroup. We study the EPE and CRS theoretically and show that they can solve many of the problems with existing metrics. We introduce a total of eight algorithms for the Cox subgroup discovery problem. The main algorithm is able to take advantage of both the EPE and the CRS, allowing us to give theoretical correctness results for this algorithm in a well-specified setting. We evaluate all of the proposed methods empirically on both synthetic and real data. The experiments confirm our theory, showing that our contributions allow for the recovery of a ground-truth subgroup in well-specified cases, as well as leading to better model fit compared to naively fitting the Cox model to the whole dataset in practical settings. Lastly, we conduct a case study on jet engine simulation data from NASA. The discovered subgroups uncover known nonlinearities/homogeneity in the data, and which suggest design choices which have been mirrored in practice.

Method: Developed an optimized Taylor-based algorithm with rigorous error analysis and dynamic selection of Taylor order and scaling factor to minimize computation under error tolerance constraints.

Result: Significant acceleration and high numerical stability compared to state-of-the-art implementations, making it highly efficient for large-scale generative modeling.

Conclusion: The proposed Taylor-based method establishes itself as a superior alternative to traditional Padé approximants for matrix exponential computation in generative AI workflows.

Abstract: The matrix exponential is a fundamental operator in scientific computing and system simulation, with applications ranging from control theory and quantum mechanics to modern generative machine learning. While Padé approximants combined with scaling and squaring have long served as the standard, recent Taylor-based methods, which utilize polynomial evaluation schemes that surpass the classical Paterson–Stockmeyer technique, offer superior accuracy and reduced computational complexity. This paper presents an optimized Taylor-based algorithm for the matrix exponential, specifically designed for the high-throughput requirements of generative AI flows. We provide a rigorous error analysis and develop a dynamic selection strategy for the Taylor order and scaling factor to minimize computational effort under a prescribed error tolerance. Extensive numerical experiments demonstrate that our approach provides significant acceleration and maintains high numerical stability compared to existing state-of-the-art implementations. These results establish the proposed method as a highly efficient tool for large-scale generative modeling.

Method: Applied the deep symbolic regression algorithm SEGVAE (Symbolic Expression Generator Variational Autoencoder) to determine properties of two-dimensional materials with defects, comparing with state-of-the-art graph neural network-based methods.

Result: SEGVAE achieved comparable or even identical outcomes to graph neural network methods, demonstrating that symbolic regression can match the performance of black-box neural networks while providing interpretable models.

Conclusion: Symbolic regression methods like SEGVAE are applicable and valuable in natural sciences, offering interpretable analytical models that can compete with state-of-the-art neural network approaches for materials science applications.

Abstract: Machine learning models have become firmly established across all scientific fields. Extracting features from data and making inferences based on them with neural network models often yields high accuracy; however, this approach has several drawbacks. Symbolic regression is a powerful technique for discovering analytical equations that describe data, providing interpretable and generalizable models capable of predicting unseen data. Symbolic regression methods have gained new momentum with the advancement of neural network technologies and offer several advantages, the main one being the interpretability of results. In this work, we examined the application of the deep symbolic regression algorithm SEGVAE to determine the properties of two-dimensional materials with defects. Comparing the results with state-of-the-art graph neural network-based methods shows comparable or, in some cases, even identical outcomes. We also discuss the applicability of this class of methods in natural sciences.

Method: A dual-specialist ensemble framework with two predictive streams: (1) Environmental Specialist using Graph Neural Network (GNN) to model community as directed contagion graph weighted by physics-informed convection, radiation, and ember probabilities, enriched with Google AlphaEarth Foundation embeddings; (2) Structural Specialist using XGBoost to isolate granular asset-level resilience. The ensemble synthesizes these through logistic stacking.

Result: Applied to 2025 Eaton Fire, the framework reveals critical dichotomy: GNN shows neighborhood-scale environmental pressure dominates structural features in propagation pathways, while XGBoost identifies eaves as primary micro-scale ingress vector. Ensemble achieves robust classification and generates diagnostic risk topology.

Conclusion: The framework enables decision-makers to move beyond binary loss prediction and precisely target mitigation - prioritizing vegetation management for high-connectivity clusters and structural hardening for architecturally vulnerable nodes, operationalizing a proactive, data-driven approach to community resilience.

Abstract: As wildfires increasingly evolve into urban conflagrations, traditional risk models that treat structures as isolated assets fail to capture the non-linear contagion dynamics characteristic of the wildland urban interface (WUI). This research bridges the gap between mechanistic physics and data driven learning by establishing a novel dual specialist ensemble framework that disentangles vulnerability into two distinct vectors, environmental contagion and structural fragility. The architecture integrates two specialized predictive streams, an environmental specialist, implemented as a graph neural network (GNN) that operationalizes the community as a directed contagion graph weighted by physics informed convection, radiation, and ember probabilities, and enriched with high dimensional Google AlphaEarth Foundation embeddings, and a Structural Specialist, implemented via XGBoost to isolate granular asset level resilience. Applied to the 2025 Eaton Fire, the framework reveals a critical dichotomy in risk drivers. The GNN demonstrates that neighborhood scale environmental pressure overwhelmingly dominates intrinsic structural features in defining propagation pathways, while the XGBoost model identifies eaves as the primary micro scale ingress vector. By synthesizing these divergent signals through logistic stacking, the ensemble achieves robust classification and generates a diagnostic risk topology. This capability empowers decision makers to move beyond binary loss prediction and precisely target mitigation prioritizing vegetation management for high connectivity clusters and structural hardening for architecturally vulnerable nodes thereby operationalizing a proactive, data driven approach to community resilience.

Method: FedMPDD encodes each client’s gradient by computing directional derivatives along multiple random vectors, compressing the gradient into a smaller message (O(m) where m « d). The server decodes aggregated information by projecting back onto the same random vectors. Using multiple projections overcomes dimension-dependent convergence limitations of single projections.

Result: Theoretical analysis shows FedMPDD converges at O(1/√K) rate, matching FedSGD performance. Experiments on benchmark datasets validate the theory. The method provides inherent privacy against gradient inversion attacks due to geometric properties of low-rank projections, with tunable privacy-utility trade-off controlled by projection count.

Conclusion: FedMPDD successfully addresses both communication efficiency and privacy in federated learning through multi-projected directional derivatives, achieving compression without sacrificing convergence while providing built-in privacy protection with controllable trade-offs.

Abstract: This paper introduces \texttt{FedMPDD} (\textbf{Fed}erated Learning via \textbf{M}ulti-\textbf{P}rojected \textbf{D}irectional \textbf{D}erivatives), a novel algorithm that simultaneously optimizes bandwidth utilization and enhances privacy in Federated Learning. The core idea of \texttt{FedMPDD} is to encode each client’s high-dimensional gradient by computing its directional derivatives along multiple random vectors. This compresses the gradient into a much smaller message, significantly reducing uplink communication costs from $\mathcal{O}(d)$ to $\mathcal{O}(m)$, where $m \ll d$. The server then decodes the aggregated information by projecting it back onto the same random vectors. Our key insight is that averaging multiple projections overcomes the dimension-dependent convergence limitations of a single projection. We provide a rigorous theoretical analysis, establishing that \texttt{FedMPDD} converges at a rate of $\mathcal{O}(1/\sqrt{K})$, matching the performance of FedSGD. Furthermore, we demonstrate that our method provides some inherent privacy against gradient inversion attacks due to the geometric properties of low-rank projections, offering a tunable privacy-utility trade-off controlled by the number of projections. Extensive experiments on benchmark datasets validate our theory and demonstrates our results.

Method: Proposes a defense system with adversarial training module within mixture-of-experts architecture. Uses nine pre-trained experts with ResNet-18 backbone. Jointly updates expert parameters and gating mechanism during end-to-end training for further optimization.

Result: The proposed defense system outperforms state-of-the-art defense systems and plain classifiers, even when those use more complex architectures than the ResNet-18 backbone.

Conclusion: The mixture-of-experts architecture with adversarial training provides effective defense against adversarial threats, demonstrating superior robustness compared to existing approaches.

Abstract: Machine learning is a powerful tool enabling full automation of a huge number of tasks without explicit programming. Despite recent progress of machine learning in different domains, these models have shown vulnerabilities when they are exposed to adversarial threats. Adversarial threats aim to hinder the machine learning models from satisfying their objectives. They can create adversarial perturbations, which are imperceptible to humans’ eyes but have the ability to cause misclassification during inference. Moreover, they can poison the training data to harm the model’s performance or they can query the model to steal its sensitive information. In this paper, we propose a defense system, which devises an adversarial training module within mixture-of-experts architecture to enhance its robustness against adversarial threats. In our proposed defense system, we use nine pre-trained experts with ResNet-18 as their backbone. During end-to-end training, the parameters of expert models and gating mechanism are jointly updated allowing further optimization of the experts. Our proposed defense system outperforms state-of-the-art defense systems and plain classifiers, which use a more complex architecture than our model’s backbone.

Method: The authors use roofline analysis to identify memory bottlenecks in BLR methods during multi-token inference. They then develop custom Triton kernels with partial fusion and memory layout optimizations specifically designed for Monarch and BLAST compression techniques to overcome these memory constraints.

Result: On memory-constrained NVIDIA GPUs (Jetson Orin Nano and A40), the optimized kernels achieve up to 3.76× speedups and 3× model size compression compared to PyTorch dense baselines with CUDA backend and compiler-level optimizations. The approach works across various models including Llama-7/1B, GPT2-S, DiT-XL/2, and ViT-B.

Conclusion: Custom memory-optimized kernels are essential for realizing the full potential of BLR compression methods in practice, especially for multi-token inference scenarios where memory bottlenecks limit performance. The proposed Triton kernels effectively address these limitations while maintaining model compression benefits.

Abstract: Recent advances in transformer-based foundation models have made them the default choice for many tasks, but their rapidly growing size makes fitting a full model on a single GPU increasingly difficult and their computational cost prohibitive. Block low-rank (BLR) compression techniques address this challenge by learning compact representations of weight matrices. While traditional low-rank (LR) methods often incur sharp accuracy drops, BLR approaches such as Monarch and BLAST can better capture the underlying structure, thus preserving accuracy while reducing computations and memory footprints. In this work, we use roofline analysis to show that, although BLR methods achieve theoretical savings and practical speedups for single-token inference, multi-token inference often becomes memory-bound in practice, increasing latency despite compiler-level optimizations in PyTorch. To address this, we introduce custom Triton kernels with partial fusion and memory layout optimizations for both Monarch and BLAST. On memory-constrained NVIDIA GPUs such as Jetson Orin Nano and A40, our kernels deliver up to $3.76\times$ speedups and $3\times$ model size compression over PyTorch dense baselines using CUDA backend and compiler-level optimizations, while supporting various models including Llama-7/1B, GPT2-S, DiT-XL/2, and ViT-B. Our code is available at https://github.com/pabillam/mem-efficient-blr .

Method: Proposes the all-pairs paired method which applies paired analysis to all pairs of LLMs and measures three noise components: prediction noise (from different answers on same question), data noise (from sampling questions), and total noise (following law of total variance). Uses millions of question-level predictions across many evaluations and settings.

Result: Two key findings: 1) Each evaluation exhibits a characteristic and highly predictable total noise level across all model pairs, 2) Paired prediction noise typically exceeds paired data noise, meaning reducing prediction noise through averaging can significantly increase statistical power.

Conclusion: The findings enable practitioners to assess significance without custom testing and to detect much smaller effects in controlled experiments, providing practical tools for more reliable LLM evaluation.

Abstract: Separating signal from noise is central to experimental science. Applying well-established statistical method effectively to LLM evals requires consideration of their unique noise characteristics. We clearly define and measure three types of noise: prediction noise from generating different answers on a given question, data noise from sampling questions, and their combined total noise following the law of total variance. To emphasize relative comparisons and gain statistical power, we propose the all-pairs paired method, which applies the paired analysis to all pairs of LLMs and measures all the noise components based on millions of question-level predictions across many evals and settings. These measurements revealed clear patterns. First, each eval exhibits a characteristic and highly predictable total noise level across all model pairs. Second, paired prediction noise typically exceeds paired data noise, which means reducing prediction noise by averaging can significantly increase statistical power. These findings enable practitioners to assess significance without custom testing and to detect much smaller effects in controlled experiments.

Method: Models gradient-based training as discrete-time dynamical system, formulates poisoning robustness as safety verification problem. Uses barrier certificates from control theory, parameterized as neural networks trained on poisoned trajectories. Derives PAC bounds via scenario convex program.

Result: Certifies non-trivial perturbation budgets on MNIST, SVHN, and CIFAR-10. Framework is model-agnostic, requires no prior knowledge of attack or contamination level. First unified framework with formal guarantees for both training and test-time attacks.

Conclusion: Provides principled formal robustness certification framework for data poisoning attacks with theoretical guarantees, addressing limitations of existing defenses and enabling reliable deployment in safety-critical applications.

Abstract: The increasing use of machine learning in safety-critical domains amplifies the risk of adversarial threats, especially data poisoning attacks that corrupt training data to degrade performance or induce unsafe behavior. Most existing defenses lack formal guarantees or rely on restrictive assumptions about the model class, attack type, extent of poisoning, or point-wise certification, limiting their practical reliability. This paper introduces a principled formal robustness certification framework that models gradient-based training as a discrete-time dynamical system (dt-DS) and formulates poisoning robustness as a formal safety verification problem. By adapting the concept of barrier certificates (BCs) from control theory, we introduce sufficient conditions to certify a robust radius ensuring that the terminal model remains safe under worst-case ${\ell}_p$-norm based poisoning. To make this practical, we parameterize BCs as neural networks trained on finite sets of poisoned trajectories. We further derive probably approximately correct (PAC) bounds by solving a scenario convex program (SCP), which yields a confidence lower bound on the certified robustness radius generalizing beyond the training set. Importantly, our framework also extends to certification against test-time attacks, making it the first unified framework to provide formal guarantees in both training and test-time attack settings. Experiments on MNIST, SVHN, and CIFAR-10 show that our approach certifies non-trivial perturbation budgets while being model-agnostic and requiring no prior knowledge of the attack or contamination level.

Method: Three-level approach: 1) Heterogeneous GNN with attention as baseline, 2) FlowKANet replacing MLPs with KAN layers (KAMP-Attn), 3) Distillation into symbolic surrogate models via block-wise regression.

Result: KAN layers reduce trainable parameters while maintaining competitive performance; symbolic surrogates eliminate trainable weights while preserving graph dependencies.

Conclusion: KAN layers offer favorable efficiency-accuracy trade-off; symbolic surrogates enable lightweight deployment and enhanced transparency for network flow prediction.

Abstract: Accurate prediction of flow delay is essential for optimizing and managing modern communication networks. We investigate three levels of modeling for this task. First, we implement a heterogeneous GNN with attention-based message passing, establishing a strong neural baseline. Second, we propose FlowKANet in which Kolmogorov-Arnold Networks replace standard MLP layers, reducing trainable parameters while maintaining competitive predictive performance. FlowKANet integrates KAMP-Attn (Kolmogorov-Arnold Message Passing with Attention), embedding KAN operators directly into message-passing and attention computation. Finally, we distill the model into symbolic surrogate models using block-wise regression, producing closed-form equations that eliminate trainable weights while preserving graph-structured dependencies. The results show that KAN layers provide a favorable trade-off between efficiency and accuracy and that symbolic surrogates emphasize the potential for lightweight deployment and enhanced transparency.

Method: The thesis proposes time-efficient methods for both red team (vulnerability identification) and blue team (mitigation) approaches within the adversarial robustness framework, though specific techniques are not detailed in the abstract.

Result: The abstract doesn’t provide specific results, but indicates the thesis successfully develops methods that overcome computational limitations of existing approaches.

Conclusion: The thesis contributes efficient methods for adversarial robustness evaluation and enhancement, making red-blue team approaches more practical for large-scale DNNs.

Abstract: With deep neural networks (DNNs) increasingly embedded in modern society, ensuring their safety has become a critical and urgent issue. In response, substantial efforts have been dedicated to the red-blue adversarial framework, where the red team focuses on identifying vulnerabilities in DNNs and the blue team on mitigating them. However, existing approaches from both teams remain computationally intensive, constraining their applicability to large-scale models. To overcome this limitation, this thesis endeavours to provide time-efficient methods for the evaluation and enhancement of adversarial robustness in DNNs.

Method: Two-stage approach: 1) Fix U-Net bottleneck as Clustering-friendly Middle Layer (CML), 2) Use Optimal Timestep Search (OTS) to find clustering-optimal timestep t*. Extract bottleneck features at t* and map via lightweight residual network. Optimize DEC-style KL self-training with adaptive graph and entropy regularization. Add denoising-consistency branch at random timesteps for stability.

Result: DiEC achieves competitive clustering performance on multiple standard benchmarks, demonstrating effectiveness of exploiting diffusion model representation trajectories for clustering.

Conclusion: Diffusion models’ internal activations across layers and timesteps contain valuable clustering information. The proposed two-stage search and regularization approach effectively extracts cluster-friendly representations from pretrained diffusion U-Nets for unsupervised clustering.

Abstract: Deep clustering hinges on learning representations that are inherently clusterable. However, using a single encoder to produce a fixed embedding ignores the representation trajectory formed by a pretrained diffusion model across network hierarchies and noise timesteps, where clusterability varies substantially. We propose DiEC (Diffusion Embedded Clustering), which performs unsupervised clustering by directly reading internal activations from a pretrained diffusion U-Net. DiEC formulates representation selection as a two-dimensional search over layer x timestep, and exploits a weak-coupling property to decompose it into two stages. Specifically, we first fix the U-Net bottleneck layer as the Clustering-friendly Middle Layer (CML), and then use Optimal Timestep Search (OTS) to identify the clustering-optimal timestep (t*). During training, we extract bottleneck features at the fixed t* and obtain clustering representations via a lightweight residual mapping. We optimize a DEC-style KL self-training objective, augmented with adaptive graph regularization and entropy regularization to strengthen cluster structures. In parallel, we introduce a denoising-consistency branch at random timesteps to stabilize the representations and preserve generative consistency. Experiments show that DiEC achieves competitive clustering performance on multiple standard benchmarks.

Method: Two-stage framework: 1) Create dataset with problem-agnostic graph features and hardness classifications, train ML classifiers to map features to hardness. 2) Use association rule mining (FP-Growth) to explain predictions, plus regression models to predict computation times.

Result: Excellent performance on 3287 maximum clique instances: weighted F1 score 0.9921, minority-class F1 0.878, ROC-AUC 0.9083 using only 3 graph features. Best association rule had 0.8829 support for hard instances with 87.64% accuracy. Best regression model achieved 5.12% RMSE and R² 0.991.

Conclusion: GCO-HPIF effectively predicts and explains computational hardness of combinatorial optimization problems, demonstrating strong performance on maximum clique problems and providing interpretable insights through association rules.

Abstract: This study introduces GCO-HPIF, a general machine-learning-based framework to predict and explain the computational hardness of combinatorial optimization problems that can be represented on graphs. The framework consists of two stages. In the first stage, a dataset is created comprising problem-agnostic graph features and hardness classifications of problem instances. Machine-learning-based classification algorithms are trained to map graph features to hardness categories. In the second stage, the framework explains the predictions using an association rule mining algorithm. Additionally, machine-learning-based regression models are trained to predict algorithmic computation times. The GCO-HPIF framework was applied to a dataset of 3287 maximum clique problem instances compiled from the COLLAB, IMDB, and TWITTER graph datasets using five state-of-the-art algorithms, namely three exact branch-and-bound-based algorithms (Gurobi, CliSAT, and MOMC) and two graph-neural-network-based algorithms (EGN and HGS). The framework demonstrated excellent performance in predicting instance hardness, achieving a weighted F1 score of 0.9921, a minority-class F1 score of 0.878, and an ROC-AUC score of 0.9083 using only three graph features. The best association rule found by the FP-Growth algorithm for explaining the hardness predictions had a support of 0.8829 for hard instances and an overall accuracy of 87.64 percent, underscoring the framework’s usefulness for both prediction and explanation. Furthermore, the best-performing regression model for predicting computation times achieved a percentage RMSE of 5.12 and an R2 value of 0.991.

Method: RevFFN introduces a memory-efficient fine-tuning paradigm for mixture of experts (MoE) LLMs using carefully designed reversible Transformer blocks. These blocks allow reconstruction of layer input activations from outputs during backpropagation, eliminating the need to store most intermediate activations in memory while preserving MoE architecture expressive capacity.

Result: The approach significantly reduces peak memory consumption for full parameter fine-tuning, enabling efficient full fine-tuning on a single consumer-grade or server-grade GPU without requiring multi-GPU memory or CPU offloading solutions.

Conclusion: RevFFN provides a practical solution for memory-efficient full parameter fine-tuning of MoE LLMs, overcoming the memory bottleneck of traditional fine-tuning approaches and making large-scale LLM adaptation more accessible with standard hardware.

Abstract: Full parameter fine tuning is a key technique for adapting large language models (LLMs) to downstream tasks, but it incurs substantial memory overhead due to the need to cache extensive intermediate activations for backpropagation. This bottleneck makes full fine tuning of contemporary large scale LLMs challenging in practice. Existing distributed training frameworks such as DeepSpeed alleviate this issue using techniques like ZeRO and FSDP, which rely on multi GPU memory or CPU offloading, but often require additional hardware resources and reduce training speed. We introduce RevFFN, a memory efficient fine tuning paradigm for mixture of experts (MoE) LLMs. RevFFN employs carefully designed reversible Transformer blocks that allow reconstruction of layer input activations from outputs during backpropagation, eliminating the need to store most intermediate activations in memory. While preserving the expressive capacity of MoE architectures, this approach significantly reduces peak memory consumption for full parameter fine tuning. As a result, RevFFN enables efficient full fine tuning on a single consumer grade or server grade GPU.

Method: Blends seasonal time-series models with tree-based learners for multivariate demand forecasting, segment-level price elasticity, and churn propensity. Uses Monte Carlo scenario testing to map risk envelopes and solves constrained optimization with business guardrails on customer experience, margin floors, and allowable churn.

Result: Validated across heterogeneous SaaS portfolios, consistently outperforms static tiers and uniform uplifts by reallocating price moves toward segments with higher willingness-to-pay while protecting price-sensitive cohorts.

Conclusion: The framework serves as a strategy playbook for shifting from flat to dynamic pricing, aligning pricing with CLV and MRR targets, and embedding ethical guardrails to enable durable growth without eroding customer trust, with real-time recalibration capabilities via modular APIs.

Abstract: This paper presents a marketing analytics framework that operationalizes subscription pricing as a dynamic, guardrailed decision system, uniting multivariate demand forecasting, segment-level price elasticity, and churn propensity to optimize revenue, margin, and retention. The approach blends seasonal time-series models with tree-based learners, runs Monte Carlo scenario tests to map risk envelopes, and solves a constrained optimization that enforces business guardrails on customer experience, margin floors, and allowable churn. Validated across heterogeneous SaaS portfolios, the method consistently outperforms static tiers and uniform uplifts by reallocating price moves toward segments with higher willingness-to-pay while protecting price-sensitive cohorts. The system is designed for real-time recalibration via modular APIs and includes model explainability for governance and compliance. Managerially, the framework functions as a strategy playbook that clarifies when to shift from flat to dynamic pricing, how to align pricing with CLV and MRR targets, and how to embed ethical guardrails, enabling durable growth without eroding customer trust.

Method: Created open-source multi-fidelity aerodynamic dataset (2448 flow snapshots across 272 geometries) using VLM and RANS solvers. Conducted scaling study with MF-VortexNet GNN surrogate using six training datasets (40-1280 snapshots) and models with 0.1-2.4M parameters under fixed training budget.

Result: Test error decreases with data size with power-law exponent of -0.6122, indicating efficient data utilization. Optimal sampling density is ~8 samples per dimension in d-dimensional design space. Larger models show improved data utilization efficiency.

Conclusion: Established empirical scaling law for GNN-based aerodynamic surrogate models, providing guidance for dataset size requirements and revealing trade-off between dataset generation cost and model training budget.

Abstract: Data-driven surrogate models are increasingly adopted to accelerate vehicle design. However, open-source multi-fidelity datasets and empirical guidelines linking dataset size to model performance remain limited. This study investigates the relationship between training data size and prediction accuracy for a graph neural network (GNN) based surrogate model for aerodynamic field prediction. We release an open-source, multi-fidelity aerodynamic dataset for double-delta wings, comprising 2448 flow snapshots across 272 geometries evaluated at angles of attack from 11 (degree) to 19 (degree) at Ma=0.3 using both Vortex Lattice Method (VLM) and Reynolds-Averaged Navier-Stokes (RANS) solvers. The geometries are generated using a nested Saltelli sampling scheme to support future dataset expansion and variance-based sensitivity analysis. Using this dataset, we conduct a preliminary empirical scaling study of the MF-VortexNet surrogate by constructing six training datasets with sizes ranging from 40 to 1280 snapshots and training models with 0.1 to 2.4 million parameters under a fixed training budget. We find that the test error decreases with data size with a power-law exponent of -0.6122, indicating efficient data utilization. Based on this scaling law, we estimate that the optimal sampling density is approximately eight samples per dimension in a d-dimensional design space. The results also suggest improved data utilization efficiency for larger surrogate models, implying a potential trade-off between dataset generation cost and model training budget.

Method: Uses Gaussian process operator learning to construct flexible functional representations directly on function space, independent of specific equations or discretizations. Incorporates physical priors through prior mean or kernel design.

Result: Achieves equal or better performance than existing approximations like local-potential approximation. Can handle non-constant fields, making it suitable for studying complex field configurations such as instantons.

Conclusion: The Gaussian process operator learning framework provides a flexible, equation-independent approach for solving functional renormalization group equations that outperforms traditional approximations and enables study of more complex field configurations.

Abstract: We present an operator learning framework for solving non-perturbative functional renormalization group equations, which are integro-differential equations defined on functionals. Our proposed approach uses Gaussian process operator learning to construct a flexible functional representation formulated directly on function space, making it independent of a particular equation or discretization. Our method is flexible, and can apply to a broad range of functional differential equations while still allowing for the incorporation of physical priors in either the prior mean or the kernel design. We demonstrate the performance of our method on several relevant equations, such as the Wetterich and Wilson–Polchinski equations, showing that it achieves equal or better performance than existing approximations such as the local-potential approximation, while being significantly more flexible. In particular, our method can handle non-constant fields, making it promising for the study of more complex field configurations, such as instantons.

Method: ReACT-Drug uses ESM-2 protein embeddings to identify similar proteins from PDB for a given target, decomposes known drug ligands into fragments to initialize search space, then employs PPO agent guiding ChemBERTa-encoded molecules through reaction-template-based transformations to generate novel compounds.

Result: Generates de novo drug candidates with competitive binding affinities and high synthetic accessibility, ensuring 100% chemical validity and novelty per MOSES benchmarking. The framework demonstrates integration of structural biology, deep representation learning, and chemical synthesis rules.

Conclusion: ReACT-Drug highlights the potential of integrating structural biology, deep representation learning, and chemical synthesis rules to automate and accelerate rational drug design through a target-agnostic RL approach.

Abstract: De novo drug design is a crucial component of modern drug development, yet navigating the vast chemical space to find synthetically accessible, high-affinity candidates remains a significant challenge. Reinforcement Learning (RL) enhances this process by enabling multi-objective optimization and exploration of novel chemical space - capabilities that traditional supervised learning methods lack. In this work, we introduce \textbf{ReACT-Drug}, a fully integrated, target-agnostic molecular design framework based on Reinforcement Learning. Unlike models requiring target-specific fine-tuning, ReACT-Drug utilizes a generalist approach by leveraging ESM-2 protein embeddings to identify similar proteins for a given target from a knowledge base such as Protein Data Base (PDB). Thereafter, the known drug ligands corresponding to such proteins are decomposed to initialize a fragment-based search space, biasing the agent towards biologically relevant subspaces. For each such fragment, the pipeline employs a Proximal Policy Optimization (PPO) agent guiding a ChemBERTa-encoded molecule through a dynamic action space of chemically valid, reaction-template-based transformations. This results in the generation of \textit{de novo} drug candidates with competitive binding affinities and high synthetic accessibility, while ensuring 100% chemical validity and novelty as per MOSES benchmarking. This architecture highlights the potential of integrating structural biology, deep representation learning, and chemical synthesis rules to automate and accelerate rational drug design. The dataset and code are available at https://github.com/YadunandanRaman/ReACT-Drug/.

Method: The authors design a prediction task where a crucial latent variable is hidden in relevant image data instead of tabular features, using a synthetic dataset for property insurance. They compare agentic AI systems that generate generic analytics code against methods that incorporate domain-specific insights.

Result: Experiments show that agentic AI relying on generic analytics workflows fails to perform well, while human experts can identify the important hidden variable using domain knowledge. This demonstrates a key limitation of current agentic AI systems.

Conclusion: Current agentic AI for data science has limitations in recognizing and incorporating domain knowledge, highlighting the need for future research to develop AI systems that can better leverage domain-specific insights like human experts.

Abstract: Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) have significantly automated data science workflows, but a fundamental question persists: Can these agentic AI systems truly match the performance of human data scientists who routinely leverage domain-specific knowledge? We explore this question by designing a prediction task where a crucial latent variable is hidden in relevant image data instead of tabular features. As a result, agentic AI that generates generic codes for modeling tabular data cannot perform well, while human experts could identify the important hidden variable using domain knowledge. We demonstrate this idea with a synthetic dataset for property insurance. Our experiments show that agentic AI that relies on generic analytics workflow falls short of methods that use domain-specific insights. This highlights a key limitation of the current agentic AI for data science and underscores the need for future research to develop agentic AI systems that can better recognize and incorporate domain knowledge.

Method: Analyze a two-layer ReLU denoising autoencoder theoretically, then validate findings on real-world unconditional and text-to-image diffusion models. Propose representation-based memorization detection and training-free editing via representation steering.

Result: Memorization corresponds to storing raw training samples in weights with localized “spiky” representations, while generalization captures local data statistics with “balanced” representations. These structures emerge in practical deep generative models.

Conclusion: Learning good representations is central to novel and meaningful generative modeling, with representation structures providing insights into memorization vs generalization and enabling practical applications for detection and control.

Abstract: Diffusion models excel at generating high-quality, diverse samples, yet they risk memorizing training data when overfit to the training objective. We analyze the distinctions between memorization and generalization in diffusion models through the lens of representation learning. By investigating a two-layer ReLU denoising autoencoder (DAE), we prove that (i) memorization corresponds to the model storing raw training samples in the learned weights for encoding and decoding, yielding localized “spiky” representations, whereas (ii) generalization arises when the model captures local data statistics, producing “balanced” representations. Furthermore, we validate these theoretical findings on real-world unconditional and text-to-image diffusion models, demonstrating that the same representation structures emerge in deep generative models with significant practical implications. Building on these insights, we propose a representation-based method for detecting memorization and a training-free editing technique that allows precise control via representation steering. Together, our results highlight that learning good representations is central to novel and meaningful generative modeling.

Method: GLiBRL uses generalised linear models with learnable basis functions to enable efficient learning of transition and reward models. It provides fully tractable marginal likelihood and Bayesian inference on task parameters and model noises.

Result: On MetaWorld ML10/45 benchmarks, GLiBRL improves VariBAD’s success rate by up to 2.7x. It outperforms other deep BRL/Meta-RL methods (MAML, RL2, SDVT, TrMRL, ECET) with low-variance and consistent performance.

Conclusion: GLiBRL addresses limitations of existing deep BRL methods by providing tractable Bayesian inference and efficient model learning, achieving superior performance on challenging meta-RL benchmarks.

Abstract: Bayesian Reinforcement Learning (BRL) provides a framework for generalisation of Reinforcement Learning (RL) problems from its use of Bayesian task parameters in the transition and reward models. However, classical BRL methods assume known forms of transition and reward models, reducing their applicability in real-world problems. As a result, recent deep BRL methods have started to incorporate model learning, though the use of neural networks directly on the joint data and task parameters requires optimising the Evidence Lower Bound (ELBO). ELBOs are difficult to optimise and may result in indistinctive task parameters, hence compromised BRL policies. To this end, we introduce a novel deep BRL method, Generalised Linear Models in Deep Bayesian RL with Learnable Basis Functions (GLiBRL), that enables efficient and accurate learning of transition and reward models, with fully tractable marginal likelihood and Bayesian inference on task parameters and model noises. On challenging MetaWorld ML10/45 benchmarks, GLiBRL improves the success rate of one of the state-of-the-art deep BRL methods, VariBAD, by up to 2.7x. Comparing against representative or recent deep BRL / Meta-RL methods, such as MAML, RL2, SDVT, TrMRL and ECET, GLiBRL also demonstrates its low-variance and decent performance consistently.

Method: CoSeNet uses a four-layer architecture (input, formatting, re-scaling, segmentation) with overlapping techniques and pre-trained ML algorithms. It optimizes re-scaling parameters using a heuristic algorithm with Window Difference-based fitness metric.

Result: The model produces a binary noise-free matrix representing optimal segmentation with segmentation points, achieving better performance than previous approaches while balancing efficiency, memory, and speed.

Conclusion: CoSeNet provides an effective, robust, and generalizable solution for correlated segment identification in noisy matrices, offering optimized parameter tuning and practical deployment advantages for various applications.

Abstract: In this paper, we propose a novel approach for the optimal identification of correlated segments in noisy correlation matrices. The proposed model is known as CoSeNet (Correlation Seg-mentation Network) and is based on a four-layer algorithmic architecture that includes several processing layers: input, formatting, re-scaling, and segmentation layer. The proposed model can effectively identify correlated segments in such matrices, better than previous approaches for similar problems. Internally, the proposed model utilizes an overlapping technique and uses pre-trained Machine Learning (ML) algorithms, which makes it robust and generalizable. CoSeNet approach also includes a method that optimizes the parameters of the re-scaling layer using a heuristic algorithm and fitness based on a Window Difference-based metric. The output of the model is a binary noise-free matrix representing optimal segmentation as well as its seg-mentation points and can be used in a variety of applications, obtaining compromise solutions between efficiency, memory, and speed of the proposed deployment model.

Method: Competitive Swiss-System Dynamics (CSD) framework with multi-round sequential contests where models are dynamically paired based on win-loss records across curated benchmarks, using Monte Carlo Simulation (100,000 iterations) to compute Expected Win Score and Failure Sensitivity Analysis with parameterized elimination to profile risk appetite.

Result: CSD provides more nuanced and context-aware rankings than traditional aggregate scoring and static pairwise models, distinguishing between robust generalists and aggressive specialists through risk profiling.

Conclusion: CSD represents a vital step towards risk-informed, next-generation LLM evaluation by addressing limitations of current static evaluation methods through dynamic competitive assessment.

Abstract: The rapid proliferation of Large Language Models (LLMs) and diverse specialized benchmarks necessitates a shift from fragmented, task-specific metrics to a holistic, competitive ranking system that effectively aggregates performance across multiple ability dimensions. Primarily using static scoring, current evaluation methods are fundamentally limited. They struggle to determine the proper mix ratio across diverse benchmarks, and critically, they fail to capture a model’s dynamic competitive fitness or its vulnerability when confronted with sequential, high-stakes tasks. To address this, we introduce the novel Competitive Swiss-System Dynamics (CSD) framework. CSD simulates a multi-round, sequential contest where models are dynamically paired across a curated sequence of benchmarks based on their accumulated win-loss record. And Monte Carlo Simulation ($N=100,000$ iterations) is used to approximate the statistically robust Expected Win Score ($E[S_m]$), which eliminates the noise of random pairing and early-round luck. Furthermore, we implement a Failure Sensitivity Analysis by parameterizing the per-round elimination quantity ($T_k$), which allows us to profile models based on their risk appetite–distinguishing between robust generalists and aggressive specialists. We demonstrate that CSD provides a more nuanced and context-aware ranking than traditional aggregate scoring and static pairwise models, representing a vital step towards risk-informed, next-generation LLM evaluation.

Method: Derived Neural Feature Dynamics (NFD) for ResNets with single-layer residual blocks in the joint infinite-width and infinite-depth limit. Studied two-layer residual blocks to understand feature collapse, then proposed a depth-aware learning-rate correction.

Result: NFD explains when scaling-law trends persist and when diminishing returns occur. Reveals that gradient-independence assumption becomes valid at infinite depth under 1/sqrt(depth) scaling. Shows feature-learning collapse in first internal layer of two-layer blocks at large depth.

Conclusion: The proposed depth-aware learning-rate correction counteracts feature collapse and restores depth-wise hyperparameter transfer, enabling better performance in deeper ResNets and providing a structural understanding of scaling limitations.

Abstract: The empirical success of deep learning is often attributed to scaling laws that predict consistent gains as model, data, and compute grow; however, large models can exhibit training instability and diminishing returns, suggesting that scaling laws describe what success looks like but not when and why scaling succeeds or fails. A central obstacle is the lack of a rigorous understanding of feature learning at large depth. While muP characterizes feature-learning dynamics in the infinite-width limit and enables hyperparameter transfer across width, its depth extension (depth-muP) breaks down for residual blocks with more than one internal layer. We derive Neural Feature Dynamics (NFD) for ResNets with single-layer residual blocks, characterizing feature learning via a coupled forward-backward stochastic system in the joint infinite-width and infinite-depth limit. In this regime, NFD identifies when scaling-law trends persist and explains diminishing returns. It also reveals a vanishing mechanism induced by the 1/sqrt(depth) residual scaling under which the gradient-independence assumption (GIA), known to fail during training at finite depth, becomes provably valid again at infinite depth, yielding an analytically tractable regime for end-to-end feature learning. Motivated by this insight, we study two-layer residual blocks and show that the same mechanism causes feature-learning collapse in the first internal layer at large depth, providing a structural explanation for the empirical failure of depth-muP. Based on this diagnosis, we propose a depth-aware learning-rate correction that counteracts the collapse and empirically restores depth-wise hyperparameter transfer, yielding stronger performance in deeper ResNets.

Method: Builds shared encoding structure for unified representation of monitoring indicators; employs state fusion mechanism for trend changes and local disturbances; introduces cross-task structural propagation module to model dependencies among nodes; incorporates dynamic adjustment mechanism for non-stationary behaviors.

Result: Superior performance on several error metrics compared to multiple models; provides more accurate representations of future states under different operating conditions; verified through hyperparameter sensitivity, environmental sensitivity, and data sensitivity analyses.

Conclusion: The unified forecasting framework offers reliable predictive capability for high-dimensional, multi-task, strongly dynamic environments in cloud native systems and provides essential technical support for intelligent backend management.

Abstract: This study proposes a unified forecasting framework for high-dimensional multi-task time series to meet the prediction demands of cloud native backend systems operating under highly dynamic loads, coupled metrics, and parallel tasks. The method builds a shared encoding structure to represent diverse monitoring indicators in a unified manner and employs a state fusion mechanism to capture trend changes and local disturbances across different time scales. A cross-task structural propagation module is introduced to model potential dependencies among nodes, enabling the model to understand complex structural patterns formed by resource contention, link interactions, and changes in service topology. To enhance adaptability to non-stationary behaviors, the framework incorporates a dynamic adjustment mechanism that automatically regulates internal feature flows according to system state changes, ensuring stable predictions in the presence of sudden load shifts, topology drift, and resource jitter. The experimental evaluation compares multiple models across various metrics and verifies the effectiveness of the framework through analyses of hyperparameter sensitivity, environmental sensitivity, and data sensitivity. The results show that the proposed method achieves superior performance on several error metrics and provides more accurate representations of future states under different operating conditions. Overall, the unified forecasting framework offers reliable predictive capability for high-dimensional, multi-task, and strongly dynamic environments in cloud native systems and provides essential technical support for intelligent backend management.

Method: Analyze single-layer Transformers from dynamical systems perspective, interpreting causal self-attention as linear, history-dependent recurrence. Investigate through linear and nonlinear case studies to identify operational regimes.

Result: For linear systems: softmax attention’s convexity constraint fundamentally restricts representable dynamics, causing oversmoothing in oscillatory settings. For nonlinear systems under partial observability: attention acts as adaptive delay-embedding mechanism enabling effective state reconstruction with sufficient temporal context and latent dimensionality.

Conclusion: The analysis bridges empirical observations with classical dynamical systems theory, providing insight into when and why Transformers succeed or fail as models of dynamical systems, helping understand their representational capabilities and limitations.

Abstract: Transformers are increasingly adopted for modeling and forecasting time-series, yet their internal mechanisms remain poorly understood from a dynamical systems perspective. In contrast to classical autoregressive and state-space models, which benefit from well-established theoretical foundations, Transformer architectures are typically treated as black boxes. This gap becomes particularly relevant as attention-based models are considered for general-purpose or zero-shot forecasting across diverse dynamical regimes. In this work, we do not propose a new forecasting model, but instead investigate the representational capabilities and limitations of single-layer Transformers when applied to dynamical data. Building on a dynamical systems perspective we interpret causal self-attention as a linear, history-dependent recurrence and analyze how it processes temporal information. Through a series of linear and nonlinear case studies, we identify distinct operational regimes. For linear systems, we show that the convexity constraint imposed by softmax attention fundamentally restricts the class of dynamics that can be represented, leading to oversmoothing in oscillatory settings. For nonlinear systems under partial observability, attention instead acts as an adaptive delay-embedding mechanism, enabling effective state reconstruction when sufficient temporal context and latent dimensionality are available. These results help bridge empirical observations with classical dynamical systems theory, providing insight into when and why Transformers succeed or fail as models of dynamical systems.

Method: STLDM is a diffusion-based model that learns latent representations end-to-end using both a Variational Autoencoder and a conditioning network. It decomposes the task into two stages: deterministic forecasting (handled by the conditioning network) and enhancement (performed by the latent diffusion model).

Result: Experimental results on multiple radar datasets show that STLDM achieves superior performance compared to state-of-the-art methods while also improving inference efficiency.

Conclusion: STLDM presents an effective diffusion-based architecture for precipitation nowcasting that successfully addresses limitations of both deterministic and generative approaches, offering improved accuracy and efficiency.

Abstract: Precipitation nowcasting is a critical spatio-temporal prediction task for society to prevent severe damage owing to extreme weather events. Despite the advances in this field, the complex and stochastic nature of this task still poses challenges to existing approaches. Specifically, deterministic models tend to produce blurry predictions while generative models often struggle with poor accuracy. In this paper, we present a simple yet effective model architecture termed STLDM, a diffusion-based model that learns the latent representation from end to end alongside both the Variational Autoencoder and the conditioning network. STLDM decomposes this task into two stages: a deterministic forecasting stage handled by the conditioning network, and an enhancement stage performed by the latent diffusion model. Experimental results on multiple radar datasets demonstrate that STLDM achieves superior performance compared to the state of the art, while also improving inference efficiency. The code is available in https://github.com/sqfoo/stldm_official.

Method: MODE dynamically combines coreset selection strategies by adapting criteria to training phases: emphasizing class balance early, diversity during representation learning, and uncertainty at convergence. It achieves (1-1/e)-approximation with O(n log n) complexity.

Result: MODE demonstrates competitive accuracy while providing interpretable insights into data utility evolution. Experiments show it reduces memory requirements compared to static methods.

Conclusion: Adaptive coreset selection that evolves with training phases is more effective than static methods, achieving better data efficiency, reduced memory requirements, and interpretable utility insights while maintaining competitive model performance.

Abstract: We present Mode(Multi-Objective adaptive Data Efficiency), a framework that dynamically combines coreset selection strategies based on their evolving contribution to model performance. Unlike static methods, \mode adapts selection criteria to training phases: emphasizing class balance early, diversity during representation learning, and uncertainty at convergence. We show that MODE achieves (1-1/e)-approximation with O(n \log n) complexity and demonstrates competitive accuracy while providing interpretable insights into data utility evolution. Experiments show \mode reduces memory requirements

Method: BALLAST uses lightweight online adaptation with contextual bandits, selecting from discrete timeout “arms” using efficient linear contextual bandits (LinUCB variants). It augments learning with safe exploration to cap risk during unstable periods.

Result: In reproducible discrete-event simulations with long-tail delay, loss, correlated bursts, node heterogeneity, and partition/recovery turbulence, BALLAST substantially reduces recovery time and unwritable time compared to standard randomized timeouts and common heuristics, while remaining competitive in stable LAN/WAN settings.

Conclusion: Contextual bandits provide an effective approach for adaptive election timeout selection in Raft, outperforming static heuristics in challenging network conditions while maintaining performance in stable environments.

Abstract: Randomized election timeouts are a simple and effective liveness heuristic for Raft, but they become brittle under long-tail latency, jitter, and partition recovery, where repeated split votes can inflate unavailability. This paper presents BALLAST, a lightweight online adaptation mechanism that replaces static timeout heuristics with contextual bandits. BALLAST selects from a discrete set of timeout “arms” using efficient linear contextual bandits (LinUCB variants), and augments learning with safe exploration to cap risk during unstable periods. We evaluate BALLAST on a reproducible discrete-event simulation with long-tail delay, loss, correlated bursts, node heterogeneity, and partition/recovery turbulence. Across challenging WAN regimes, BALLAST substantially reduces recovery time and unwritable time compared to standard randomized timeouts and common heuristics, while remaining competitive on stable LAN/WAN settings.

Method: U-GEPSVM extends GEPSVM framework by incorporating Universum constraints through ratio-based objective function. IU-GEPSVM enhances stability through weighted difference-based formulation providing independent control over class separation and Universum alignment.

Result: IU-GEPSVM achieves peak accuracies of 85% (O vs S) and 80% (Z vs S), with mean accuracies of 81.29% and 77.57% respectively, outperforming baseline methods on Bonn University EEG dataset.

Conclusion: The proposed Universum-enhanced classifiers effectively improve EEG signal classification performance for seizure detection, demonstrating the value of combining computational efficiency with Universum learning for handling EEG analysis challenges.

Abstract: The paper presents novel Universum-enhanced classifiers: the Universum Generalized Eigenvalue Proximal Support Vector Machine (U-GEPSVM) and the Improved U-GEPSVM (IU-GEPSVM) for EEG signal classification. Using the computational efficiency of generalized eigenvalue decomposition and the generalization benefits of Universum learning, the proposed models address critical challenges in EEG analysis: non-stationarity, low signal-to-noise ratio, and limited labeled data. U-GEPSVM extends the GEPSVM framework by incorporating Universum constraints through a ratio-based objective function, while IU-GEPSVM enhances stability through a weighted difference-based formulation that provides independent control over class separation and Universum alignment. The models are evaluated on the Bonn University EEG dataset across two binary classification tasks: (O vs S)-healthy (eyes closed) vs seizure, and (Z vs S)-healthy (eyes open) vs seizure. IU-GEPSVM achieves peak accuracies of 85% (O vs S) and 80% (Z vs S), with mean accuracies of 81.29% and 77.57% respectively, outperforming baseline methods.

Method: Introduces ICRL prompting - a multi-round framework where LLMs receive numerical scalar rewards after each response, then are prompted again with concatenated prior responses and rewards to optimize performance iteratively.

Result: ICRL prompting shows significant improvements over baselines (Self-Refine, Reflexion) on Game of 24, creative writing, ScienceWorld, and math competitions (AIME, HMMT), even when rewards are generated by the same LLM.

Conclusion: LLMs can perform in-context reinforcement learning during inference, optimizing scalar reward signals without training, representing a promising paradigm for test-time scaling and self-improvement.

Abstract: Reinforcement learning (RL) is a framework for solving sequential decision-making problems. In this work, we demonstrate that, surprisingly, RL emerges during the inference time of large language models (LLMs), a phenomenon we term in-context RL (ICRL). To reveal this capability, we introduce a simple multi-round prompting framework, we call ICRL prompting, for inference-time self-improvement. The goal of ICRL prompting is to guide LLMs to perform reinforcement learning during inference for self-improvement on a given task. After each response, the model receives numerical scalar feedback, denoted as a reward. In the next round, we prompt the LLM again together with a context that concatenates all prior responses and their associated rewards. We consistently observe that response quality improves as the context grows. In other words, the LLM can optimize scalar reward signals during inference, exhibiting behavior analogous to reinforcement learning. We evaluate ICRL prompting on Game of 24, creative writing, ScienceWorld, and Olympiad-level math competitions (AIME and HMMT), demonstrating significant improvements over baselines such as Self-Refine and Reflexion. Notably, even when the reward signals are generated by the same LLM, ICRL prompting still improves performance, highlighting a promising new paradigm for test-time scaling.

Method: Introduces a Learning Stability Profile that tracks infinitesimal responses of representations, parameters, and update mechanisms to perturbations along learning trajectories. Uses analytic and variational approaches with Lyapunov-type energy functions, extending to non-smooth systems via Clarke generalized derivatives and variational Lyapunov functionals.

Result: Proves a Fundamental Analytic Stability Theorem showing equivalence between uniform boundedness of stability signatures and existence of dissipative Lyapunov-type energy. Derives explicit stability exponents linking spectral norms, activation regularity, step sizes, and learning rates to contractivity. Recovers classical stability results as special cases and extends framework to non-smooth systems.

Conclusion: The framework provides a unified dynamical description of stability across architectures and optimization methods, clarifying how architectural and algorithmic choices jointly govern robustness and sensitivity to perturbations, and establishes foundation for extensions to continuous-time limits and geometric formulations.

Abstract: We propose a unified analytic and variational framework for studying stability in deep learning systems viewed as coupled representation-parameter dynamics. The central object is the Learning Stability Profile, which tracks the infinitesimal response of representations, parameters, and update mechanisms to perturbations along the learning trajectory. We prove a Fundamental Analytic Stability Theorem showing that uniform boundedness of these stability signatures is equivalent, up to norm equivalence, to the existence of a Lyapunov-type energy that dissipates along the learning flow. In smooth regimes, the framework yields explicit stability exponents linking spectral norms, activation regularity, step sizes, and learning rates to contractivity of the learning dynamics. Classical spectral stability results for feedforward networks, a discrete CFL-type condition for residual architectures, and parametric and temporal stability laws for stochastic gradient methods arise as direct consequences. The theory extends to non-smooth learning systems, including ReLU networks, proximal and projected updates, and stochastic subgradient flows, by replacing classical derivatives with Clarke generalized derivatives and smooth energies with variational Lyapunov functionals. The resulting framework provides a unified dynamical description of stability across architectures and optimization methods, clarifying how architectural and algorithmic choices jointly govern robustness and sensitivity to perturbations. It also provides a foundation for further extensions to continuous-time limits and geometric formulations of learning dynamics.

Method: Proposed MiST (mid-stage scientific training): data-mixing with SMILES/CIF-aware pre-processing, continued pre-training on 2.9B tokens, and supervised fine-tuning on 1B tokens. This builds symbolic competence and latent chemical knowledge needed for RL success.

Result: MiST raised latent-solvability scores by up to 1.8x on 3B and 7B models. RL then lifted top-1 accuracy from 10.9% to 63.9% on organic reaction naming, and from 40.6% to 67.4% on inorganic material generation. Similar improvements on other chemical tasks with interpretable reasoning traces.

Conclusion: Clear prerequisites for chemical reasoning training are identified (symbolic competence + latent chemical knowledge). Mid-stage training (MiST) is crucial for unlocking reasoning capabilities in LLMs for scientific domains like chemistry.

Abstract: Large Language Models can develop reasoning capabilities through online fine-tuning with rule-based rewards. However, recent studies reveal a critical constraint: reinforcement learning succeeds only when the base model already assigns non-negligible probability to correct answers – a property we term ’latent solvability’. This work investigates the emergence of chemical reasoning capabilities and what these prerequisites mean for chemistry. We identify two necessary conditions for RL-based chemical reasoning: 1) Symbolic competence, and 2) Latent chemical knowledge. We propose mid-stage scientific training (MiST): a set of mid-stage training techniques to satisfy these, including data-mixing with SMILES/CIF-aware pre-processing, continued pre-training on 2.9B tokens, and supervised fine-tuning on 1B tokens. These steps raise the latent-solvability score on 3B and 7B models by up to 1.8x, and enable RL to lift top-1 accuracy from 10.9 to 63.9% on organic reaction naming, and from 40.6 to 67.4% on inorganic material generation. Similar results are observed for other challenging chemical tasks, while producing interpretable reasoning traces. Our results define clear prerequisites for chemical reasoning training and highlight the broader role of mid-stage training in unlocking reasoning capabilities.

Method: Proposes ARS-OPT algorithm inspired by Nesterov’s Accelerated Gradient (NAG) that proactively estimates gradients with respect to future ray directions using accumulated momentum. Also introduces PARS-OPT which incorporates surrogate-model priors for enhanced gradient estimation.

Result: Theoretical analysis shows ARS-OPT enables more accurate directional updates and achieves faster, more stable optimization. Extensive experiments on ImageNet and CIFAR-10 demonstrate superiority over 13 state-of-the-art approaches in query efficiency.

Conclusion: The proposed momentum-based approach with Nesterov acceleration principles significantly improves query efficiency in hard-label black-box adversarial attacks, making such attacks more practical for deployment.

Abstract: In hard-label black-box adversarial attacks, where only the top-1 predicted label is accessible, the prohibitive query complexity poses a major obstacle to practical deployment. In this paper, we focus on optimizing a representative class of attacks that search for the optimal ray direction yielding the minimum $\ell_2$-norm perturbation required to move a benign image into the adversarial region. Inspired by Nesterov’s Accelerated Gradient (NAG), we propose a momentum-based algorithm, ARS-OPT, which proactively estimates the gradient with respect to a future ray direction inferred from accumulated momentum. We provide a theoretical analysis of its convergence behavior, showing that ARS-OPT enables more accurate directional updates and achieves faster, more stable optimization. To further accelerate convergence, we incorporate surrogate-model priors into ARS-OPT’s gradient estimation, resulting in PARS-OPT with enhanced performance. The superiority of our approach is supported by theoretical guarantees under standard assumptions. Extensive experiments on ImageNet and CIFAR-10 demonstrate that our method surpasses 13 state-of-the-art approaches in query efficiency.

Method: Three key contributions: (1) Develops a flatness-aware PAC-Bayes generalization bound for model merging with a “cross-task heterogeneity” term; (2) Frames merging as multi-teacher knowledge distillation on unlabeled data; (3) Implements SAMerging using Sharpness-Aware Minimization to find flat minima that minimize KL divergence between student and teachers.

Result: SAMerging establishes new state-of-the-art performance across vision and NLP benchmarks, demonstrating remarkable empirical performance improvements over existing methods.

Conclusion: The paper provides a principled theoretical foundation for model merging through PAC-Bayes analysis, operationalizes it via knowledge distillation framing, and demonstrates practical effectiveness with SAMerging, offering a robust alternative to heuristic approaches.

Abstract: Model merging has emerged as a lightweight alternative to joint multi-task learning (MTL), yet the generalization properties of merged models remain largely unexplored. Establishing such theoretical guarantees is non-trivial, as the merging process typically forbids access to the original training data and involves combining fine-tuned models trained on fundamentally heterogeneous data distributions. Without a principled understanding of these dynamics, current methods often rely on heuristics to approximate the optimal combination of parameters. This dependence is most critical in coefficient scaling, the weighting factors that modulate the magnitude of each fine-tuned model’s contribution to the shared parameter. However, without a principled objective to guide their selection, these methods lead to brittle performance and are highly sensitive to scaling initialization. We address this gap by (i) establishing a novel flatness-aware PAC-Bayes generalization bound specifically for the model merging setting. This analysis introduces a “cross-task heterogeneity” term that formally captures the mismatch between diverse fine-tuned model priors and the target multi-task distributions. Guided by this theoretical insight, (ii) we frame model merging as multi-teacher knowledge distillation on scarce, unlabeled data. We formally demonstrate that minimizing the student-teacher Kullback-Leibler divergence directly tightens the upper bound on the merged model’s excess risk. Guided by the flatness-aware bound derived, (iii) we operationalize this objective via SAMerging, a method that employs Sharpness-Aware Minimization (SAM) to find flat minima. Empirically, SAMerging establishes a new state of the art across vision and NLP benchmarks, achieving remarkable performance. The code is available at https://github.com/arshandalili/SAMerging.

Method: TFPI introduces a simple ThinkFree operation that explicitly discards thinking content via direct append to reduce token usage during inference. Training with ThinkFree-adapted inputs improves performance and lowers token consumption even in original slow-thinking mode.

Result: TFPI accelerates RL convergence, achieves higher performance ceiling, and yields more token-efficient reasoning models without specialized rewards or complex training designs. A 4B model trained with TFPI reached 89.0% accuracy on AIME24 and 65.5% on LiveCodeBench using less than 4K H20 hours.

Conclusion: TFPI is an effective adaptation to RLVR that bridges CoT distillation and standard RLVR, enabling more efficient training and inference while maintaining or improving performance across various benchmarks.

Abstract: Reinforcement Learning with Verifiable Reward (RLVR) effectively solves complex tasks but demands extremely long context lengths during training, leading to substantial computational costs. While multi-stage training can partially mitigate this, starting with overly short contexts often causes irreversible performance degradation, ultimately failing to reduce overall training compute significantly. In this paper, we introduce Thinking-Free Policy Initialization (TFPI), a simple yet effective adaptation to RLVR that bridges long Chain-of-Thought (CoT) distillation and standard RLVR. TFPI employs a simple ThinkFree operation, explicitly discarding the thinking content via a direct append, to reduce token usage during inference. Training with ThinkFree-adapted inputs improves performance and lowers token consumption, even in the original slow-thinking mode. Extensive experiments across various benchmarks have shown that TFPI accelerates RL convergence, achieves a higher performance ceiling, and yields more token-efficient reasoning models without specialized rewards or complex training designs. With TFPI only, we train a 4B model to reach 89.0% accuracy on AIME24 and 65.5% on LiveCodeBench using less than 4K H20 hours.

Method: Analyzed TCGA-LAML RNA-seq data with WGCNA to identify 20 biomarkers, modeled structures with AlphaFold3, mapped druggable hotspots with DOGSiteScorer, and developed reaction-first evolutionary metaheuristic algorithm for fragment-based ligand assembly with multi-objective optimization.

Result: Generated structurally unique chemical entities with drug-like properties (QED scores 0.5-0.7), identified high-confidence candidates like Ligand L1 with binding free energy of -6.571 kcal/mol against A08A96 biomarker through ADMET profiling and molecular docking.

Conclusion: Integrating systems biology with metaheuristic molecular assembly can produce pharmacologically viable, patient-tailored drug leads, offering a scalable blueprint for precision oncology in AML and other cancers.

Abstract: Acute Myeloid Leukemia (AML) remains a clinical challenge due to its extreme molecular heterogeneity and high relapse rates. While precision medicine has introduced mutation-specific therapies, many patients still lack effective, personalized options. This paper presents a novel, end-to-end computational framework that bridges the gap between patient-specific transcriptomics and de novo drug discovery. By analyzing bulk RNA sequencing data from the TCGA-LAML cohort, the study utilized Weighted Gene Co-expression Network Analysis (WGCNA) to prioritize 20 high-value biomarkers, including metabolic transporters like HK3 and immune-modulatory receptors such as SIGLEC9. The physical structures of these targets were modeled using AlphaFold3, and druggable hotspots were quantitatively mapped via the DOGSiteScorer engine. Then developed a novel, reaction-first evolutionary metaheuristic algorithm as well as multi-objective optimization programming that assembles novel ligands from fragment libraries, guided by spatial alignment to these identified hotspots. The generative model produced structurally unique chemical entities with a strong bias toward drug-like space, as evidenced by QED scores peaking between 0.5 and 0.7. Validation through ADMET profiling and SwissDock molecular docking identified high-confidence candidates, such as Ligand L1, which achieved a binding free energy of -6.571 kcal/mol against the A08A96 biomarker. These results demonstrate that integrating systems biology with metaheuristic molecular assembly can produce pharmacologically viable, patient tailored leads, offering a scalable blueprint for precision oncology in AML and beyond

Method: Learns a local update operator conditioned on neural shape attributes, integrating naturally with neural surface representations, trained once on a single representative shape and generalizing across variations.

Result: Slightly outperforms CPM while remaining reasonably close to FEM, delivers first end-to-end pipeline for solving surface PDEs on both neural and classical surface representations.

Conclusion: Enables accurate, fast inference without explicit meshing or per-instance optimization while preserving differentiability, bridging the gap between neural representations and PDE solving.

Abstract: Solving partial differential equations (PDEs) on shapes underpins many shape analysis and engineering tasks; yet, prevailing PDE solvers operate on polygonal/triangle meshes while modern 3D assets increasingly live as neural representations. This mismatch leaves no suitable method to solve surface PDEs directly within the neural domain, forcing explicit mesh extraction or per-instance residual training, preventing end-to-end workflows. We present a novel, mesh-free formulation that learns a local update operator conditioned on neural (local) shape attributes, enabling surface PDEs to be solved directly where the (neural) data lives. The operator integrates naturally with prevalent neural surface representations, is trained once on a single representative shape, and generalizes across shape and topology variations, enabling accurate, fast inference without explicit meshing or per-instance optimization while preserving differentiability. Across analytic benchmarks (heat equation and Poisson solve on sphere) and real neural assets across different representations, our method slightly outperforms CPM while remaining reasonably close to FEM, and, to our knowledge, delivers the first end-to-end pipeline that solves surface PDEs on both neural and classical surface representations. Code will be released on acceptance.

Method: Theoretical analysis of binary classification with a classifier that approximates the optimal Bayes classifier, plus empirical studies on both theoretical setup and practical deep classifiers on benchmark datasets with varying noise levels, training sizes, and class distributions.

Result: Proved that for any finite number of training samples, there exists pre-classification processing that improves classification accuracy. Empirical studies confirm these theoretical findings in practical settings.

Conclusion: While the data processing inequality holds for optimal Bayes classifiers, pre-processing can benefit practical classifiers with finite training data, with gains influenced by class separation, training size, and class balance.

Abstract: The data processing inequality is an information-theoretic principle stating that the information content of a signal cannot be increased by processing the observations. In particular, it suggests that there is no benefit in enhancing the signal or encoding it before addressing a classification problem. This assertion can be proven to be true for the case of the optimal Bayes classifier. However, in practice, it is common to perform “low-level” tasks before “high-level” downstream tasks despite the overwhelming capabilities of modern deep neural networks. In this paper, we aim to understand when and why low-level processing can be beneficial for classification. We present a comprehensive theoretical study of a binary classification setup, where we consider a classifier that is tightly connected to the optimal Bayes classifier and converges to it as the number of training samples increases. We prove that for any finite number of training samples, there exists a pre-classification processing that improves the classification accuracy. We also explore the effect of class separation, training set size, and class balance on the relative gain from this procedure. We support our theory with an empirical investigation of the theoretical setup. Finally, we conduct an empirical study where we investigate the effect of denoising and encoding on the performance of practical deep classifiers on benchmark datasets. Specifically, we vary the size and class distribution of the training set, and the noise level, and demonstrate trends that are consistent with our theoretical results.

Method: A 3-phase block-level WSD training scheme: 1) progressive increasing block-size in block diffusion (warm-up), 2) large-scale full-sequence diffusion (stable), and 3) reverting to compact-size block diffusion (decay). Post-training alignment with SFT and DPO creates MoE variants.

Result: Created LLaDA2.0-mini (16B) and LLaDA2.0-flash (100B) instruction-tuned MoE variants optimized for practical deployment, delivering superior performance and efficiency at frontier scale with parallel decoding.

Conclusion: LLaDA2.0 establishes a new paradigm for frontier-scale discrete diffusion LLM deployment through systematic conversion from auto-regressive models, enabling knowledge inheritance and efficient parallel decoding while being open-sourced.

Abstract: This paper presents LLaDA2.0 – a tuple of discrete diffusion large language models (dLLM) scaling up to 100B total parameters through systematic conversion from auto-regressive (AR) models – establishing a new paradigm for frontier-scale deployment. Instead of costly training from scratch, LLaDA2.0 upholds knowledge inheritance, progressive adaption and efficiency-aware design principle, and seamless converts a pre-trained AR model into dLLM with a novel 3-phase block-level WSD based training scheme: progressive increasing block-size in block diffusion (warm-up), large-scale full-sequence diffusion (stable) and reverting back to compact-size block diffusion (decay). Along with post-training alignment with SFT and DPO, we obtain LLaDA2.0-mini (16B) and LLaDA2.0-flash (100B), two instruction-tuned Mixture-of-Experts (MoE) variants optimized for practical deployment. By preserving the advantages of parallel decoding, these models deliver superior performance and efficiency at the frontier scale. Both models were open-sourced.

Method: Proposes TimeBridge framework using diffusion bridges to learn paths between a chosen prior and data distribution, with specialized prior designs including data/time-dependent priors for unconditional generation and scale-preserving priors for conditional generation.

Result: Experiments show TimeBridge with data-driven priors outperforms standard diffusion models on time series generation tasks.

Conclusion: TimeBridge provides a flexible framework for time series synthesis with tailored priors that better capture time series properties, achieving superior performance over conventional diffusion approaches.

Abstract: Time series generation is widely used in real-world applications such as simulation, data augmentation, and hypothesis testing. Recently, diffusion models have emerged as the de facto approach to time series generation, enabling diverse synthesis scenarios. However, the fixed standard-Gaussian diffusion prior may be ill-suited for time series data, which exhibit properties such as temporal order and fixed time points. In this paper, we propose TimeBridge, a framework that flexibly synthesizes time series data by using diffusion bridges to learn paths between a chosen prior and the data distribution. We then explore several prior designs tailored to time series synthesis. Our framework covers (i) data- and time-dependent priors for unconditional generation and (ii) scale-preserving priors for conditional generation. Experiments show that our framework with data-driven priors outperforms standard diffusion models on time series generation.

Method: Formalize the problem as reinforcement learning, derive conditions for image information sufficiency, introduce scalable benchmark, use overcomplete sparse coding for image representation, and provide theoretical justification.

Result: Reinforcement learning provides computationally efficient method for finding optimal policies with natural images when encoded as efficient representations; sparse codes enable solving control tasks orders of magnitude larger than with complete codes.

Conclusion: Deep learning is not necessary for efficient optimal control with natural images; sparse coding representations combined with reinforcement learning enable scalable solutions to large-scale control problems.

Abstract: Optimal control and sequential decision making are widely used in many complex tasks. Optimal control over a sequence of natural images is a first step towards understanding the role of vision in control. Here, we formalize this problem as a reinforcement learning task, and derive general conditions under which an image includes enough information to implement an optimal policy. Reinforcement learning is shown to provide a computationally efficient method for finding optimal policies when natural images are encoded into “efficient” image representations. This is demonstrated by introducing a new reinforcement learning benchmark that easily scales to large numbers of states and long horizons. In particular, by representing each image as an overcomplete sparse code, we are able to efficiently solve an optimal control task that is orders of magnitude larger than those tasks solvable using complete codes. Theoretical justification for this behaviour is provided. This work also demonstrates that deep learning is not necessary for efficient optimal control with natural images.

Method: Develops Stochastic Pessimistic-Optimistic Thresholding (SPOT), a model-based primal-dual algorithm using Growing-Window estimator to sample from environment interactions and estimate stochastic thresholds, handling multiple constraints against unknown thresholds.

Result: Proves SPOT achieves sublinear regret and constraint violation: $\tilde{\mathcal{O}}(\sqrt{T})$ reward regret with $\tilde{\mathcal{O}}(\sqrt{T})$ constraint violation over T episodes, matching performance of approaches with fixed known thresholds.

Conclusion: SPOT is the first RL algorithm with theoretical guarantees for uncertain environments where thresholds are unknown, enabling safe learning comparable to fixed-threshold approaches while handling stochastic threshold uncertainty.

Abstract: This paper studies constrained Markov decision processes (CMDPs) with constraints against stochastic thresholds, aiming at safety of reinforcement learning in unknown and uncertain environments. We leverage a Growing-Window estimator sampling from interactions with the uncertain environment to estimate the thresholds, based on which we design Stochastic Pessimistic-Optimistic Thresholding (SPOT), a novel model-based primal-dual algorithm for multiple constraints against stochastic thresholds. SPOT enables reinforcement learning under both pessimistic and optimistic threshold settings. We prove that our algorithm achieves sublinear regret and constraint violation; i.e., a reward regret of $\tilde{\mathcal{O}}(\sqrt{T})$ while allowing an $\tilde{\mathcal{O}}(\sqrt{T})$ constraint violation over $T$ episodes. The theoretical guarantees show that our algorithm achieves performance comparable to that of an approach relying on fixed and clear thresholds. To the best of our knowledge, SPOT is the first reinforcement learning algorithm that realises theoretical guaranteed performance in an uncertain environment where even thresholds are unknown.

Method: Proposes a weighted aggregation framework where weights are assigned to each prediction set based on their contribution. Derives a procedure for weighted aggregation that maintains finite-sample validity even when weights depend on the data. Generalizes to settings where weights are learned, such as mixture-of-experts.

Result: Achieves tighter coverage bounds that interpolate between 1-2α (combined models) and 1-α (individual model) depending on weight distribution. Maintains finite-sample validity with data-dependent weights. Demonstrates adaptive coverage in mixture-of-experts experiments.

Conclusion: The weighted aggregation framework provides flexible control over prediction set aggregation while maintaining strong coverage guarantees, making it broadly applicable to settings with learned weights like mixture-of-experts.

Abstract: Conformal prediction quantifies the uncertainty of machine learning models by augmenting point predictions with valid prediction sets. For complex scenarios involving multiple trials, models, or data sources, conformal prediction sets can be aggregated to create a prediction set that captures the overall uncertainty, often improving precision. However, aggregating multiple prediction sets with individual $1-α$ coverage inevitably weakens the overall guarantee, typically resulting in $1-2α$ worst-case coverage. In this work, we propose a framework for the weighted aggregation of prediction sets, where weights are assigned to each prediction set based on their contribution. Our framework offers flexible control over how the sets are aggregated, achieving tighter coverage bounds that interpolate between the $1-2α$ guarantee of the combined models and the $1-α$ guarantee of an individual model depending on the distribution of weights. Importantly, our framework generalizes to data-dependent weights, as we derive a procedure for weighted aggregation that maintains finite-sample validity even when the weights depend on the data. This extension makes our framework broadly applicable to settings where weights are learned, such as mixture-of-experts (MoE), and we demonstrate through experiments in the MoE setting that our methods achieve adaptive coverage.

Method: Represent AI models as DAGs, reformulate splitting as minimum s-t cut problem, propose fast DAG-based algorithm using maximum flow method, and block-wise algorithm for block-structured models.

Result: Algorithms find optimal splitting within milliseconds and reduce training delay by 24.62%-38.95% compared to state-of-the-art benchmarks in dynamic edge networks.

Conclusion: The proposed DAG-based approach provides optimal and efficient model splitting for split learning, significantly improving training performance in edge computing environments.

Abstract: Split learning (SL) has emerged as a computationally efficient approach for artificial intelligence (AI) model training, which can alleviate device-side computational workloads. However, complex AI model architectures pose high computational complexity to obtain the optimal model splitting. In this paper, we represent an arbitrary AI model as a directed acyclic graph (DAG), and then reformulate the optimal model splitting problem as a minimum s-t cut search problem. To solve the problem, we propose a fast DAG-based model splitting algorithm, which restructures the DAG to enable the optimal model splitting identification via a maximum flow method. Theoretical analysis indicates that the proposed algorithm is optimal. Furthermore, considering AI models with block structures, we propose a block-wise model splitting algorithm to reduce computational complexity. The algorithm abstracts each block, i.e., a component consisting of multiple layers, into a single vertex, thereby obtaining the optimal model splitting via a simplified DAG. Extensive experimental results demonstrate that the proposed algorithms can determine the optimal model splitting within milliseconds, as well as reduce training delay by 24.62%-38.95% in dynamic edge networks as compared to the state-of-the-art benchmarks.

Method: Introduces stochastic activations that randomly select between SILU or RELU via Bernoulli draws during training. Two strategies: (1) Use stochastic activations during pre-training, fine-tune with RELU for sparse inference; (2) Apply stochastic activations for sequence generation to increase diversity.

Result: Stochastic activations outperform training from scratch with RELU, reduce inference FLOPs with significant CPU/GPU speedup, and provide higher diversity in text generation with only slightly inferior performance to SILU with temperature sampling.

Conclusion: Stochastic activations effectively address RELU’s optimization limitations, enable efficient sparse inference, and provide an alternative approach for increasing text generation diversity in language models.

Abstract: We introduce stochastic activations. This novel strategy randomly selects between several non-linear functions in the feed-forward layer of a large language model. In particular, we choose between SILU or RELU depending on a Bernoulli draw. This strategy circumvents the optimization problem associated with RELU, namely, the constant shape for negative inputs that prevents the gradient flow. We leverage this strategy in two ways: (1) We use stochastic activations during pre-training and fine-tune the model with RELU, which is used at inference time to provide sparse latent vectors. This reduces the inference FLOPs and translates into a significant speedup on CPU and GPU. This leads to better results than training from scratch with the RELU activation function. (2) We evaluate stochastic activations for sequence generation. This strategy performs reasonably well: it has higher diversity and has only slightly inferior performance to the best deterministic non-linearity, SILU, combined with temperature sampling. This provides an alternative way to increase the diversity of generated text.

Method: Designs a sparse projection method that simultaneously projects groups of vectors with explicit average sparsity control using Hoyer measure, with linear complexity. Also proposes a weighted ℓ₁ norm generalization.

Result: Method achieves linear computational complexity. In ResNet50 pruning, produces sparse models with significantly higher accuracy at corresponding sparsity levels. In nonnegative matrix factorization, yields competitive reconstruction errors against state-of-the-art algorithms.

Conclusion: The proposed sparse projection method effectively controls average sparsity for vector sets with linear complexity, demonstrating superior performance in deep learning pruning and competitive results in matrix factorization tasks.

Abstract: We design a new sparse projection method for a set of vectors that guarantees a desired average sparsity level measured leveraging the popular Hoyer measure (an affine function of the ratio of the $\ell_1$ and $\ell_2$ norms). Existing approaches either project each vector individually or require the use of a regularization parameter which implicitly maps to the average $\ell_0$-measure of sparsity. Instead, in our approach we set the sparsity level for the whole set explicitly and simultaneously project a group of vectors with the sparsity level of each vector tuned automatically. We show that the computational complexity of our projection operator is linear in the size of the problem. Additionally, we propose a generalization of this projection by replacing the $\ell_1$ norm by its weighted version. We showcase the efficacy of our approach in both supervised and unsupervised learning tasks on image datasets including CIFAR10 and ImageNet. In deep neural network pruning, the sparse models produced by our method on ResNet50 have significantly higher accuracies at corresponding sparsity values compared to existing competitors. In nonnegative matrix factorization, our approach yields competitive reconstruction errors against state-of-the-art algorithms.

Method: The paper formalizes verifier unreliability as a stochastic reward channel with asymmetric noise rates (FP rate ρ₀ and FN rate ρ₁). It proposes two lightweight corrections: (1) backward correction yielding unbiased surrogate reward and unbiased policy-gradient estimator, and (2) forward correction that reweights score-function terms to align with clean gradient direction (requires only FN rate). Both are implemented as hooks in group relative policy optimization pipeline. An appeals mechanism with lightweight LLM verifier estimates FN rate online.

Result: Both corrections improve RLVR for math reasoning under synthetic and real verifier noise. The forward variant is more stable under heavier noise. The appeals mechanism with lightweight LLM verifier for online FN rate estimation further improves performance.

Conclusion: The paper successfully addresses verifier unreliability in RLVR through formal modeling of noise channels and lightweight corrections. The forward correction shows particular robustness to heavy noise, and online FN rate estimation via appeals mechanism provides additional performance gains, making RLVR more practical with imperfect automated verifiers.

Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) replaces costly human labeling with automated verifiers. To reduce verifier hacking, many RLVR systems binarize rewards to ${0,1}$, but imperfect verifiers inevitably introduce \emph{false negatives} (rejecting correct answers) and \emph{false positives} (accepting incorrect ones). We formalize verifier unreliability as a stochastic reward channel with asymmetric noise rates $ρ_0$ and $ρ_1$ – the FP rate and the FN rate, respectively. From this abstraction we derive two lightweight corrections: (i) a \emph{backward} correction that yields an unbiased surrogate reward and thus an unbiased policy-gradient estimator in expectation, and (ii) a \emph{forward} correction that reweights score-function terms so the expected update aligns with the clean gradient direction and requires only the FN rate. We implement both as lightweight hooks in a group relative policy optimization pipeline, both corrections improve RLVR for math reasoning under synthetic and real verifier noise, with the forward variant being more stable under heavier noise. Finally, an appeals mechanism with a lightweight LLM verifier estimates the FN rate online and further improves performance.

Method: DATTA uses a domain-diversity discriminator to recognize domain patterns, domain-diversity adaptive batch normalization to combine source and test-time statistics, and domain-diversity adaptive fine-tuning to resolve gradient conflicts.

Result: Extensive experiments show DATTA significantly outperforms state-of-the-art methods by up to 13% improvement.

Conclusion: DATTA is the first approach to successfully handle TTA under dynamic domain shift data streams, providing robust adaptation to changing domain distributions in real-world scenarios.

Abstract: Test-Time Adaptation (TTA) addresses domain shifts between training and testing. However, existing methods assume a homogeneous target domain (e.g., single domain) at any given time. They fail to handle the dynamic nature of real-world data, where single-domain and multiple-domain distributions change over time. We identify that performance drops in multiple-domain scenarios are caused by batch normalization errors and gradient conflicts, which hinder adaptation. To solve these challenges, we propose Domain Diversity Adaptive Test-Time Adaptation (DATTA), the first approach to handle TTA under dynamic domain shift data streams. It is guided by a novel domain-diversity score. DATTA has three key components: a domain-diversity discriminator to recognize single- and multiple-domain patterns, domain-diversity adaptive batch normalization to combine source and test-time statistics, and domain-diversity adaptive fine-tuning to resolve gradient conflicts. Extensive experiments show that DATTA significantly outperforms state-of-the-art methods by up to 13%. Code is available at https://github.com/DYW77/DATTA.

Method: Develops H-embedding, an online-computable transferability-aware task embedding based on information theoretic measures. Uses this embedding to guide a hypernet framework that learns task-conditioned model weights for continual learning tasks.

Result: Extensive evaluations on CIFAR-100, ImageNet-R, and DomainNet benchmarks show prominent performance compared to various baseline and SOTA approaches, demonstrating strong potential in capturing and utilizing intrinsic task relationships.

Conclusion: The proposed H-embedding guided hypernet framework effectively enhances continual learning by leveraging inter-task relationships through transferability-aware embeddings, offering practical advantages with low-dimensional storage and efficient end-to-end training.

Abstract: Continual learning (CL) has been a critical topic in contemporary deep neural network applications, where higher levels of both forward and backward transfer are desirable for an effective CL performance. Existing CL strategies primarily focus on task models, either by regularizing model updates or by separating task-specific and shared components, while often overlooking the potential of leveraging inter-task relationships to enhance transfer. To address this gap, we propose a transferability-aware task embedding, termed H-embedding, and construct a hypernet framework under its guidance to learn task-conditioned model weights for CL tasks. Specifically, H-embedding is derived from an information theoretic measure of transferability and is designed to be online and easy to compute. Our method is also characterized by notable practicality, requiring only the storage of a low-dimensional task embedding per task and supporting efficient end-to-end training. Extensive evaluations on benchmarks including CIFAR-100, ImageNet-R, and DomainNet show that our framework performs prominently compared to various baseline and SOTA approaches, demonstrating strong potential in capturing and utilizing intrinsic task relationships. Our code is publicly available at https://github.com/viki760/H-embedding-Guided-Hypernet.

Method: PEARL dynamically allocates ranks for LoRA components during CL training by leveraging reference task weights and adaptively determining rank based on current task’s proximity to reference weights in parameter space.

Result: Outperforms all considered baselines by a large margin across three vision architectures (ResNet, Separable Convolutional Network, Vision Transformer) and multiple CL scenarios.

Conclusion: PEARL provides an effective rehearsal-free CL framework with dynamic rank allocation that addresses LoRA’s rank sensitivity while maintaining parameter efficiency and preventing catastrophic forgetting.

Abstract: Catastrophic forgetting has remained a critical challenge for deep neural networks in Continual Learning (CL) as it undermines consolidated knowledge when learning new tasks. Parameter efficient fine tuning CL techniques are gaining traction for their effectiveness in addressing catastrophic forgetting with a lightweight training schedule while avoiding degradation of consolidated knowledge in pre-trained models. However, low rank adapters (LoRA) in these approaches are highly sensitive to rank selection which can lead to sub-optimal resource allocation and performance. To this end, we introduce PEARL, a rehearsal-free CL framework that entails dynamic rank allocation for LoRA components during CL training. Specifically, PEARL leverages reference task weights and adaptively determines the rank of task-specific LoRA components based on the current tasks’ proximity to reference task weights in parameter space. To demonstrate the versatility of PEARL, we evaluate it across three vision architectures (ResNet, Separable Convolutional Network and Vision Transformer) and a multitude of CL scenarios, and show that PEARL outperforms all considered baselines by a large margin.

Method: Two AI-based approaches: 1) Deep Symbolic Regression for fully interpretable models, and 2) Kolmogorov-Arnold Networks offering two levels of interpretability. Both automate model formulation, evaluation, and refinement. Evaluated on two synthetic and two real-world datasets.

Result: Kolmogorov-Arnold Networks achieve R² close to 1 with minimal prediction error. Deep Symbolic Regression produces compact models with moderate accuracy. Automated methods outperform traditional approaches, achieving up to 75% reduction in prediction errors.

Conclusion: The proposed automated AI methods provide accurate and explainable solutions for path loss modeling, potentially increasing efficiency in discovering next-generation propagation models for 5G+ wireless systems.

Abstract: Modeling propagation is the cornerstone for designing and optimizing next-generation wireless systems, with a particular emphasis on 5G and beyond era. Traditional modeling methods have long relied on statistic-based techniques to characterize propagation behavior across different environments. With the expansion of wireless communication systems, there is a growing demand for methods that guarantee the accuracy and interpretability of modeling. Artificial intelligence (AI)-based techniques, in particular, are increasingly being adopted to overcome this challenge, although the interpretability is not assured with most of these methods. Inspired by recent advancements in AI, this paper proposes a novel approach that accelerates the discovery of path loss models while maintaining interpretability. The proposed method automates the formulation, evaluation, and refinement of the model, facilitating the discovery of the model. We examine two techniques: one based on Deep Symbolic Regression, offering full interpretability, and the second based on Kolmogorov-Arnold Networks, providing two levels of interpretability. Both approaches are evaluated on two synthetic and two real-world datasets. Our results show that Kolmogorov-Arnold Networks achieve the coefficient of determination value R^2 close to 1 with minimal prediction error, while Deep Symbolic Regression generates compact models with moderate accuracy. Moreover, on the selected examples, we demonstrate that automated methods outperform traditional methods, achieving up to 75% reduction in prediction errors, offering accurate and explainable solutions with potential to increase the efficiency of discovering next-generation path loss models.

Method: Alternating Gradient Flows (AGF) approximates gradient flow dynamics as a two-step alternating process: (1) maximizing a utility function over dormant neurons to activate them, and (2) minimizing a cost function over active neurons. The framework begins with all neurons dormant and activates one per iteration, triggering feature acquisition.

Result: AGF successfully quantifies the order, timing, and magnitude of loss drops, matching experimental results across various architectures. It unifies existing saddle-to-saddle analyses, proves convergence to gradient flow in diagonal linear networks, and provides the first complete characterization of training dynamics in quadratic networks for modular addition, revealing Fourier feature learning by coefficient magnitude.

Conclusion: AGF offers a promising framework for understanding feature learning in neural networks, explaining staircase loss patterns, feature acquisition order, and providing theoretical insights across multiple network architectures from a unified perspective.

Abstract: What features neural networks learn, and how, remains an open question. In this paper, we introduce Alternating Gradient Flows (AGF), an algorithmic framework that describes the dynamics of feature learning in two-layer networks trained from small initialization. Prior works have shown that gradient flow in this regime exhibits a staircase-like loss curve, alternating between plateaus where neurons slowly align to useful directions and sharp drops where neurons rapidly grow in norm. AGF approximates this behavior as an alternating two-step process: maximizing a utility function over dormant neurons and minimizing a cost function over active ones. AGF begins with all neurons dormant. At each iteration, a dormant neuron activates, triggering the acquisition of a feature and a drop in the loss. AGF quantifies the order, timing, and magnitude of these drops, matching experiments across several commonly studied architectures. We show that AGF unifies and extends existing saddle-to-saddle analyses in fully connected linear networks and attention-only linear transformers, where the learned features are singular modes and principal components, respectively. In diagonal linear networks, we prove AGF converges to gradient flow in the limit of vanishing initialization. Applying AGF to quadratic networks trained to perform modular addition, we give the first complete characterization of the training dynamics, revealing that networks learn Fourier features in decreasing order of coefficient magnitude. Altogether, AGF offers a promising step towards understanding feature learning in neural networks.

Method: DaSH (Dataset Selection via Hierarchies) formalizes dataset selection as selecting entire datasets from heterogeneous pools. It models utility at both dataset and group levels (e.g., collections, institutions), enabling efficient generalization from limited observations through hierarchical modeling.

Result: Across Digit-Five and DomainNet benchmarks, DaSH outperforms state-of-the-art data selection baselines by up to 26.2% in accuracy while requiring significantly fewer exploration steps. It’s robust to low-resource settings and lack of relevant datasets.

Conclusion: DaSH provides an effective solution for scalable and adaptive dataset selection in practical multi-source learning workflows, addressing the limitations of sample-level selection methods by considering hierarchical dataset structures and source variations.

Abstract: The success of modern machine learning hinges on access to high-quality training data. In many real-world scenarios, such as acquiring data from public repositories or sharing across institutions, data is naturally organized into discrete datasets that vary in relevance, quality, and utility. Selecting which repositories or institutions to search for useful datasets, and which datasets to incorporate into model training are therefore critical decisions, yet most existing methods select individual samples and treat all data as equally relevant, ignoring differences between datasets and their sources. In this work, we formalize the task of dataset selection: selecting entire datasets from a large, heterogeneous pool to improve downstream performance under resource constraints. We propose Dataset Selection via Hierarchies (DaSH), a dataset selection method that models utility at both dataset and group (e.g., collections, institutions) levels, enabling efficient generalization from limited observations. Across two public benchmarks (Digit-Five and DomainNet), DaSH outperforms state-of-the-art data selection baselines by up to 26.2% in accuracy, while requiring significantly fewer exploration steps. Ablations show DaSH is robust to low-resource settings and lack of relevant datasets, making it suitable for scalable and adaptive dataset selection in practical multi-source learning workflows.

Method: Integrates two key methods: (1) physics-informed neural networks with hyper-networks that construct an ODE solver, and (2) Wasserstein generative adversarial networks that estimate ODE parameters by capturing noisy data distributions.

Result: SiGMoID successfully quantifies data noise, estimates system parameters, and infers unobserved system components. Its effectiveness is validated through realistic experimental examples.

Conclusion: The approach demonstrates broad applicability across various domains from scientific research to engineered systems, enabling discovery of full system dynamics from imperfect data.

Abstract: System inference for nonlinear dynamic models, represented by ordinary differential equations (ODEs), remains a significant challenge in many fields, particularly when the data are noisy, sparse, or partially observable. In this paper, we propose a Simulation-based Generative Model for Imperfect Data (SiGMoID) that enables precise and robust inference for dynamic systems. The proposed approach integrates two key methods: (1) physics-informed neural networks with hyper-networks that constructs an ODE solver, and (2) Wasserstein generative adversarial networks that estimates ODE parameters by effectively capturing noisy data distributions. We demonstrate that SiGMoID quantifies data noise, estimates system parameters, and infers unobserved system components. Its effectiveness is validated validated through realistic experimental examples, showcasing its broad applicability in various domains, from scientific research to engineered systems, and enabling the discovery of full system dynamics.

Method: AdaMuon incorporates two key mechanisms: 1) element-wise second momentum estimator applied to orthogonalized update directions, and 2) sign-stabilized orthogonal update where momentum is sign-transformed before orthogonalization. It also uses RMS-aligned rescaling to match Adam’s update magnitude for compatibility with existing learning rate schedules.

Result: Experiments show AdaMuon maintains stability while achieving over 40% better training efficiency than Adam in large-scale scenarios, without requiring additional tuning of learning rate schedules.

Conclusion: AdaMuon successfully combines element-wise adaptivity with orthogonal updates to create a stable and efficient optimizer for large-scale neural network training that outperforms Adam while maintaining compatibility with existing learning rate schedules.

Abstract: We propose AdaMuon, a novel optimizer that combines element-wise adaptivity with orthogonal updates for large-scale neural network training. AdaMuon incorporates two tightly coupled mechanisms: (1) an element-wise second momentum estimator applied to orthogonalized update directions, and (2) a sign-stabilized orthogonal update, where the momentum is first sign-transformed before orthogonalization. These two components jointly enable variance-adaptive scaling while maintaining stable update geometry. In addition, AdaMuon employs an RMS-aligned rescaling strategy to match the root-mean-square update magnitude to Adam, allowing direct reuse of existing learning rate schedules without extra tuning. Experiments demonstrate that AdaMuon not only maintains stability but can surpass Adam by more than 40% training efficiency in large-scale scenarios.

Method: Benchmarked Pareto-based MOBO strategy (Expected Hypervolume Improvement - EHVI) against fixed-weight scalarized baseline (Expected Improvement - EI) using identical Gaussian Process surrogates and molecular representations in tightly controlled setup across three molecular optimization tasks.

Result: EHVI consistently outperformed scalarized EI in terms of Pareto front coverage, convergence speed, and chemical diversity. Even strong deterministic scalarization variants underperform in low-data regimes.

Conclusion: Pareto-aware acquisition (EHVI) offers practical advantages over scalarized approaches in de novo molecular optimization, especially with limited evaluation budgets and nontrivial trade-offs.

Abstract: Multi-objective Bayesian optimization (MOBO) provides a principled framework for navigating trade-offs in molecular design. However, its empirical advantages over scalarized alternatives remain underexplored. We benchmark a simple Pareto-based MOBO strategy - Expected Hypervolume Improvement (EHVI) - against a simple fixed-weight scalarized baseline using Expected Improvement (EI), under a tightly controlled setup with identical Gaussian Process surrogates and molecular representations. Across three molecular optimization tasks, EHVI consistently outperforms scalarized EI in terms of Pareto front coverage, convergence speed, and chemical diversity. While scalarization encompasses flexible variants - including random or adaptive schemes - our results show that even strong deterministic instantiations can underperform in low-data regimes. These findings offer concrete evidence for the practical advantages of Pareto-aware acquisition in de novo molecular optimization, especially when evaluation budgets are limited and trade-offs are nontrivial.

Method: Evaluated LASSO logistic regression, random forest, XGBoost, and neural networks using clinical feature subsets (including top 10 SHAP-ranked features). Applied equal opportunity postprocessing to reduce disparities in true positive rates across racial/ethnic subgroups. Used large EHR database with training (59,492), validation (24,198), and testing (25,188) datasets.

Result: Selected LASSO logistic regression with top 10 features for interpretability and comparable performance. Before fairness adjustment: AUROC 0.84, accuracy 78%, sensitivity 72%, specificity 79%, F1-score 0.617. After equal opportunity postprocessing: accuracy increased to 81%, specificity to 94%, but sensitivity decreased to 41% and F1-score to 0.515, reflecting fairness trade-off.

Conclusion: MASER (MASLD Static EHR Risk Prediction) demonstrates that interpretable models can achieve competitive performance (AUROC 0.836, accuracy 77.6%) comparable to ensemble and tree-based models while balancing predictive performance and fairness in diverse patient populations.

Abstract: Background: Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD) affects ~33% of U.S. adults and is the most common chronic liver disease. Although often asymptomatic, progression can lead to cirrhosis. Early detection is important, as lifestyle interventions can prevent disease progression. We developed a fair, rigorous, and reproducible MASLD prediction model and compared it to prior methods using a large electronic health record database. Methods: We evaluated LASSO logistic regression, random forest, XGBoost, and a neural network for MASLD prediction using clinical feature subsets, including the top 10 SHAP-ranked features. To reduce disparities in true positive rates across racial and ethnic subgroups, we applied an equal opportunity postprocessing method. Results: This study included 59,492 patients in the training data, 24,198 in the validating data, and 25,188 in the testing data. The LASSO logistic regression model with the top 10 features was selected for its interpretability and comparable performance. Before fairness adjustment, the model achieved AUROC of 0.84, accuracy of 78%, sensitivity of 72%, specificity of 79%, and F1-score of 0.617. After equal opportunity postprocessing, accuracy modestly increased to 81% and specificity to 94%, while sensitivity decreased to 41% and F1-score to 0.515, reflecting the fairness trade-off. Conclusions: We developed the MASER prediction model (MASLD Static EHR Risk Prediction), a LASSO logistic regression model which achieved competitive performance for MASLD prediction (AUROC 0.836, accuracy 77.6%), comparable to previously reported ensemble and tree-based models. Overall, this approach demonstrates that interpretable models can achieve a balance of predictive performance and fairness in diverse patient populations.

Method: Introduces a higher-order, non-causal sequence model that outputs control signals for the residual stream activations of a base autoregressive model. This learns to compress long activation sequences into internal controllers that execute meaningful action sequences with learned termination conditions.

Result: On grid world and MuJoCo-based hierarchical tasks, the higher-order model learns to compress activation sequences into controllers that execute behaviorally meaningful actions over long timescales. Internal RL enables learning from sparse rewards where standard RL fails.

Conclusion: Internal RL demonstrates benefits of latent action generation and reinforcement in autoregressive models, offering a promising approach for hierarchical RL within foundation models by enabling efficient exploration through temporally-abstract internal controllers.

Abstract: Large-scale autoregressive models pretrained on next-token prediction and finetuned with reinforcement learning (RL) have achieved unprecedented success on many problem domains. During RL, these models explore by generating new outputs, one token at a time. However, sampling actions token-by-token can result in highly inefficient learning, particularly when rewards are sparse. Here, we show that it is possible to overcome this problem by acting and exploring within the internal representations of an autoregressive model. Specifically, to discover temporally-abstract actions, we introduce a higher-order, non-causal sequence model whose outputs control the residual stream activations of a base autoregressive model. On grid world and MuJoCo-based tasks with hierarchical structure, we find that the higher-order model learns to compress long activation sequence chunks onto internal controllers. Critically, each controller executes a sequence of behaviorally meaningful actions that unfold over long timescales and are accompanied with a learned termination condition, such that composing multiple controllers over time leads to efficient exploration on novel tasks. We show that direct internal controller reinforcement, a process we term “internal RL”, enables learning from sparse rewards in cases where standard RL finetuning fails. Our results demonstrate the benefits of latent action generation and reinforcement in autoregressive models, suggesting internal RL as a promising avenue for realizing hierarchical RL within foundation models.

Method: Proposed framework incorporates physics-based constraints into prediction models. Introduces novel Physics Informed Kolmogorov Arnold Network (PIKAN) that blends universal function approximation theory with physical insights. Structural parameters collected manually or through computer vision. Dataset includes 15 real bridges augmented to 100 samples.

Result: Best model achieves R² score of 0.9603 and mean absolute error (MAE) of 10.50 units. Also developed web-based interface for parameter entry and prediction. Shows PINNs can provide reliable structural weight estimates with limited data.

Conclusion: PINNs offer reliable estimates of structural weight even with limited data and may help inform early-stage failure analysis in lightweight bridge designs. The novel PIKAN architecture successfully blends physics with neural networks for structural prediction tasks.

Abstract: Physics Informed Neural Networks (PINNs) are gaining attention for their ability to embed physical laws into deep learning models, which is particularly useful in structural engineering tasks with limited data. This paper aims to explore the use of PINNs to predict the weight of small scale spaghetti bridges, a task relevant to understanding load limits and potential failure modes in simplified structural models. Our proposed framework incorporates physics-based constraints to the prediction model for improved performance. In addition to standard PINNs, we introduce a novel architecture named Physics Informed Kolmogorov Arnold Network (PIKAN), which blends universal function approximation theory with physical insights. The structural parameters provided as input to the model are collected either manually or through computer vision methods. Our dataset includes 15 real bridges, augmented to 100 samples, and our best model achieves an $R^2$ score of 0.9603 and a mean absolute error (MAE) of 10.50 units. From applied perspective, we also provide a web based interface for parameter entry and prediction. These results show that PINNs can offer reliable estimates of structural weight, even with limited data, and may help inform early stage failure analysis in lightweight bridge designs. The complete data and code are available at https://github.com/OmerJauhar/PINNS-For-Spaghetti-Bridges.

Method: Introduces SynQuE problem formalization, establishes comprehensive benchmarks, adapts distribution/diversity-based distance measures via embedding models, and proposes LENS - a novel proxy that leverages large language model reasoning for complex tasks.

Result: SynQuE proxies correlate with real task performance across diverse tasks (sentiment analysis, Text2SQL, web navigation, image classification). LENS consistently outperforms others on complex tasks. On text-to-SQL parsing, training on top-3 synthetic datasets selected via SynQuE raises accuracy from 30.4% to 38.4% (+8.1%) compared to indiscriminate selection.

Conclusion: Establishes SynQuE as a practical framework for synthetic data selection under real-data scarcity and motivates future research on foundation model-based data characterization and fine-grained data selection.

Abstract: We introduce and formalize the Synthetic Dataset Quality Estimation (SynQuE) problem: ranking synthetic datasets by their expected real-world task performance using only limited unannotated real data. This addresses a critical and open challenge where data is scarce due to collection costs or privacy constraints. We establish the first comprehensive benchmarks for this problem by introducing and evaluating proxy metrics that choose synthetic data for training to maximize task performance on real data. We introduce the first proxy metrics for SynQuE by adapting distribution and diversity-based distance measures to our context via embedding models. To address the shortcomings of these metrics on complex planning tasks, we propose LENS, a novel proxy that leverages large language model reasoning. Our results show that SynQuE proxies correlate with real task performance across diverse tasks, including sentiment analysis, Text2SQL, web navigation, and image classification, with LENS consistently outperforming others on complex tasks by capturing nuanced characteristics. For instance, on text-to-SQL parsing, training on the top-3 synthetic datasets selected via SynQuE proxies can raise accuracy from 30.4% to 38.4 (+8.1)% on average compared to selecting data indiscriminately. This work establishes SynQuE as a practical framework for synthetic data selection under real-data scarcity and motivates future research on foundation model-based data characterization and fine-grained data selection.

Method: Integrates Kolmogorov-Arnold Networks (KANs) within fair adversarial learning framework, leveraging KANs’ adversarial robustness and interpretability. Includes theoretical insights on spline-based KAN architecture for stable adversarial optimization and adaptive fairness penalty update mechanism to balance fairness and accuracy.

Result: Empirical evidence on two real-world admissions datasets demonstrates the framework’s efficiency in achieving fairness across sensitive attributes while preserving predictive performance.

Conclusion: The proposed KAN-based fair adversarial learning framework effectively addresses bias in predictive models, achieving better fairness-accuracy trade-offs with improved interpretability for socially sensitive applications like college admissions.

Abstract: Despite recent advances in fairness-aware machine learning, predictive models often exhibit discriminatory behavior towards marginalized groups. Such unfairness might arise from biased training data, model design, or representational disparities across groups, posing significant challenges in high-stakes decision-making domains such as college admissions. While existing fair learning models aim to mitigate bias, achieving an optimal trade-off between fairness and accuracy remains a challenge. Moreover, the reliance on black-box models hinders interpretability, limiting their applicability in socially sensitive domains. To circumvent these issues, we propose integrating Kolmogorov-Arnold Networks (KANs) within a fair adversarial learning framework. Leveraging the adversarial robustness and interpretability of KANs, our approach facilitates stable adversarial learning. We derive theoretical insights into the spline-based KAN architecture that ensure stability during adversarial optimization. Additionally, an adaptive fairness penalty update mechanism is proposed to strike a balance between fairness and accuracy. We back these findings with empirical evidence on two real-world admissions datasets, demonstrating the proposed framework’s efficiency in achieving fairness across sensitive attributes while preserving predictive performance.

Method: The authors propose a common design framework for representative shortcut models that provides theoretical justification for their validity and disentangles concrete component-level choices. This framework enables systematic identification of improvements without requiring pre-training, distillation, or curriculum learning.

Result: The improved one-step model achieves state-of-the-art FID50k of 2.85 on ImageNet-256x256 under classifier-free guidance with one-step generation, and further reaches FID50k of 2.53 with 2x training steps. The model requires no pre-training, distillation, or curriculum learning.

Conclusion: The proposed framework lowers the barrier to component-level innovation in shortcut models and facilitates principled exploration of their design space, enabling systematic improvements and state-of-the-art performance in one-step diffusion models.

Abstract: Recent advances in few-step diffusion models have demonstrated their efficiency and effectiveness by shortcutting the probabilistic paths of diffusion models, especially in training one-step diffusion models from scratch (\emph{a.k.a.} shortcut models). However, their theoretical derivation and practical implementation are often closely coupled, which obscures the design space. To address this, we propose a common design framework for representative shortcut models. This framework provides theoretical justification for their validity and disentangles concrete component-level choices, thereby enabling systematic identification of improvements. With our proposed improvements, the resulting one-step model achieves a new state-of-the-art FID50k of 2.85 on ImageNet-256x256 under the classifier-free guidance setting with one step generation, and further reaches FID50k of 2.53 with 2x training steps. Remarkably, the model requires no pre-training, distillation, or curriculum learning. We believe our work lowers the barrier to component-level innovation in shortcut models and facilitates principled exploration of their design space.

Method: Adapt arguments for complex VAEs with complex latent stage; derive Fisher information metric for complex Gaussian with trivial relation matrix; propose Kähler potential derivative of complex Gaussian mixtures as efficient proxy to Fisher metric; use law of total covariance to bridge potential and metric; regularize latent space with decoder geometry and sample with weighted complex volume element.

Result: The method enables efficient computation of Fisher metric via plurisubharmonic potential function, displacing large-scale automatic differentiation burden to small scale; regularization yields consistently smoother representations and fewer semantic outliers, though at the exchange of sample variation.

Conclusion: Complex VAEs naturally admit Kähler geometric structure; the proposed Kähler potential derivative provides an efficient computational framework for Fisher information metrics while maintaining geometric faithfulness; decoder geometry regularization improves representation quality.

Abstract: It has been discovered that latent-Euclidean variational autoencoders (VAEs) admit, in various capacities, Riemannian structure. We adapt these arguments but for complex VAEs with a complex latent stage. We show that complex VAEs reveal to some level Kähler geometric structure. Our methods will be tailored for decoder geometry. We derive the Fisher information metric in the complex case under a latent complex Gaussian with trivial relation matrix. It is well known from statistical information theory that the Fisher information coincides with the Hessian of the Kullback-Leibler (KL) divergence. Thus, the metric Kähler potential relation is exactly achieved under relative entropy. We propose a Kähler potential derivative of complex Gaussian mixtures that acts as a rough proxy to the Fisher information metric while still being faithful to the underlying Kähler geometry. Computation of the metric via this potential is efficient, and through our potential, valid as a plurisubharmonic (PSH) function, large scale computational burden of automatic differentiation is displaced to small scale. Our methods leverage the law of total covariance to bridge behavior between our potential and the Fisher metric. We show that we can regularize the latent space with decoder geometry, and that we can sample in accordance with a weighted complex volume element. We demonstrate these strategies, at the exchange of sample variation, yield consistently smoother representations and fewer semantic outliers.

Method: DDRL uses forward KL divergence to anchor the policy to an off-policy data distribution, enabling robust integration of RL with standard diffusion training through reward maximization combined with diffusion loss minimization.

Result: With extensive experiments (over 1M GPU hours) and human evaluations (10k double-blind), DDRL significantly improves rewards while alleviating reward hacking in high-resolution video generation, achieving highest human preference.

Conclusion: DDRL establishes a robust and scalable paradigm for diffusion post-training by preventing reward hacking through data-regularized reinforcement learning.

Abstract: Aligning generative diffusion models with human preferences via reinforcement learning (RL) is critical yet challenging. Most existing algorithms are often vulnerable to reward hacking, such as quality degradation, over-stylization, or reduced diversity. Our analysis demonstrates that this can be attributed to the inherent limitations of their regularization, which provides unreliable penalties. We introduce Data-regularized Diffusion Reinforcement Learning (DDRL), a novel framework that uses the forward KL divergence to anchor the policy to an off-policy data distribution. Theoretically, DDRL enables robust, unbiased integration of RL with standard diffusion training. Empirically, this translates into a simple yet effective algorithm that combines reward maximization with diffusion loss minimization. With over a million GPU hours of experiments and ten thousand double-blind human evaluations, we demonstrate on high-resolution video generation tasks that DDRL significantly improves rewards while alleviating the reward hacking seen in baselines, achieving the highest human preference and establishing a robust and scalable paradigm for diffusion post-training.

Method: The paper evaluates limitations of existing interference prediction approaches, analyzing how coarse-grained methods and static models perform under different conditions.

Result: Coarse-grained methods lead to noticeable deviations in prediction accuracy, and static models degrade considerably under changing workloads.

Conclusion: Current interference prediction approaches have significant limitations that restrict their usefulness for scheduling in ML inference serving systems, highlighting the need for more accurate and adaptive methods.

Abstract: Machine learning (ML) inference serving systems can schedule requests to improve GPU utilization and to meet service level objectives (SLOs) or deadlines. However, improving GPU utilization may compromise latency-sensitive scheduling, as concurrent tasks contend for GPU resources and thereby introduce interference. Given that interference effects introduce unpredictability in scheduling, neglecting them may compromise SLO or deadline satisfaction. Nevertheless, existing interference prediction approaches remain limited in several respects, which may restrict their usefulness for scheduling. First, they are often coarse-grained, which ignores runtime co-location dynamics and thus restricts their accuracy in interference prediction. Second, they tend to use a static prediction model, which may not effectively cope with different workload characteristics. In this paper, we evaluate the potential limitations of existing interference prediction approaches, finding that coarse-grained methods can lead to noticeable deviations in prediction accuracy and that static models degrade considerably under changing workloads.

Method: Proposed CAse Prompting (CAP) framework that integrates conventional debiasing prompts with case-based reasoning. CAP guides models to learn from similar historical misprediction cases and their correct outcomes, enabling correction of biased reasoning patterns without retraining. First developed a multi-dimensional bias assessment scheme for comprehensive diagnosis.

Result: On MIMIC-IV dataset, CAP increased AUROC from 0.806 to 0.873 and AUPRC from 0.497 to 0.694, while reducing sex- and race-related disparities by over 90%. Feature reliance analysis showed highly consistent attention patterns across demographic groups (similarity scores >0.98).

Conclusion: LLMs exhibit measurable bias in ICU mortality prediction, but a carefully designed prompting framework can effectively co-optimize fairness and performance without retraining. CAP offers a transferable paradigm for equitable clinical decision support.

Abstract: Accurate mortality risk prediction for intensive care unit (ICU) patients is essential for clinical decision-making. Although large language models (LLMs) show promise in predicting outcomes from structured medical data, their predictions may exhibit demographic biases related to sex, age, and race, limiting their trustworthy use in clinical practice. Existing debiasing methods often reduce predictive performance, making it difficult to jointly optimize fairness and accuracy. In this study, we systematically examine bias in LLM-based ICU mortality prediction and propose a training-free, clinically adaptive prompting framework to simultaneously improve fairness and performance. We first develop a multi-dimensional bias assessment scheme for comprehensive model diagnosis. Building on this analysis, we introduce CAse Prompting (CAP), a novel prompting framework that integrates conventional debiasing prompts with case-based reasoning. CAP guides the model to learn from similar historical misprediction cases and their correct outcomes, enabling correction of biased reasoning patterns. Experiments on the MIMIC-IV dataset show that CAP substantially improves both predictive accuracy and fairness. CAP increases AUROC from 0.806 to 0.873 and AUPRC from 0.497 to 0.694, while reducing sex- and race-related disparities by over 90%. Feature reliance analysis further indicates highly consistent attention patterns across demographic groups, with similarity scores exceeding 0.98. These results demonstrate that LLMs exhibit measurable bias in ICU mortality prediction, and that a carefully designed prompting framework can effectively co-optimize fairness and performance without retraining, offering a transferable paradigm for equitable clinical decision support.

Method: GeoTransolver uses a Multiscale Geometry-Aware Physics Attention Transformer that replaces standard attention with GALE (Geometry-Aware Latent Embedding). It couples physics-aware self-attention on learned state slices with cross-attention to a shared geometry/global/boundary-condition context computed from multi-scale ball queries (inspired by DoMINO). The method persistently projects geometry, global and boundary condition parameters into physical state spaces to anchor latent computations to domain structure and operating regimes.

Result: GeoTransolver delivers better accuracy, improved robustness to geometry/regime shifts, and favorable data efficiency compared to benchmarks including Domino, Transolver, and AB-UPT. The paper includes ablations on DrivAerML and qualitative results such as contour plots and design trends for the best GeoTransolver models.

Conclusion: By unifying multiscale geometry-aware context with physics-based attention in a scalable transformer, GeoTransolver advances operator learning for high-fidelity surrogate modeling across complex, irregular domains and non-linear physical regimes.

Abstract: We present GeoTransolver, a Multiscale Geometry-Aware Physics Attention Transformer for CAE that replaces standard attention with GALE, coupling physics-aware self-attention on learned state slices with cross-attention to a shared geometry/global/boundary-condition context computed from multi-scale ball queries (inspired by DoMINO) and reused in every block. Implemented and released in NVIDIA PhysicsNeMo, GeoTransolver persistently projects geometry, global and boundary condition parameters into physical state spaces to anchor latent computations to domain structure and operating regimes. We benchmark GeoTransolver on DrivAerML, Luminary SHIFT-SUV, and Luminary SHIFT-Wing, comparing against Domino, Transolver (as released in PhysicsNeMo), and literature-reported AB-UPT, and evaluate drag/lift R2 and Relative L1 errors for field variables. GeoTransolver delivers better accuracy, improved robustness to geometry/regime shifts, and favorable data efficiency; we include ablations on DrivAerML and qualitative results such as contour plots and design trends for the best GeoTransolver models. By unifying multiscale geometry-aware context with physics-based attention in a scalable transformer, GeoTransolver advances operator learning for high-fidelity surrogate modeling across complex, irregular domains and non-linear physical regimes.

Method: Novel decentralized framework for optimal joint action selection that explicitly accounts for belief inconsistencies, provides probabilistic guarantees for action consistency and performance relative to open-loop multi-agent POMDP, and selectively triggers communication only when needed.

Result: Simulation results show the approach outperforms state-of-the-art algorithms.

Conclusion: The framework addresses critical challenges of belief inconsistency in multi-agent systems by providing probabilistic guarantees and efficient communication strategies, improving both coordination and safety.

Abstract: Multi-agent decision-making under uncertainty is fundamental for effective and safe autonomous operation. In many real-world scenarios, each agent maintains its own belief over the environment and must plan actions accordingly. However, most existing approaches assume that all agents have identical beliefs at planning time, implying these beliefs are conditioned on the same data. Such an assumption is often impractical due to limited communication. In reality, agents frequently operate with inconsistent beliefs, which can lead to poor coordination and suboptimal, potentially unsafe, performance. In this paper, we address this critical challenge by introducing a novel decentralized framework for optimal joint action selection that explicitly accounts for belief inconsistencies. Our approach provides probabilistic guarantees for both action consistency and performance with respect to open-loop multi-agent POMDP (which assumes all data is always communicated), and selectively triggers communication only when needed. Furthermore, we address another key aspect of whether, given a chosen joint action, the agents should share data to improve expected performance in inference. Simulation results show our approach outperforms state-of-the-art algorithms.

Method: Three key innovations: (1) on-chain DAO governance for transparent coordination and immutable logging, (2) ZKP mechanism for off-chain Shapley-based contribution measurement, and (3) hybrid on-chain/off-chain architecture that verifies ZKP-validated contributions on-chain with minimal computational overhead.

Result: Experimental results using crypto trading tasks show DAO-Agent achieves up to 99.9% reduction in verification gas costs compared to naive on-chain alternatives, with constant-time verification complexity that remains stable as coalition size increases.

Conclusion: DAO-Agent establishes a scalable foundation for agent coordination in decentralized environments by balancing transparency, fairness, privacy preservation, and cost efficiency through its innovative architecture.

Abstract: Autonomous Large Language Model (LLM)-based multi-agent systems have emerged as a promising paradigm for facilitating cross-application and cross-organization collaborations. These autonomous agents often operate in trustless environments, where centralized coordination faces significant challenges, such as the inability to ensure transparent contribution measurement and equitable incentive distribution. While blockchain is frequently proposed as a decentralized coordination platform, it inherently introduces high on-chain computation costs and risks exposing sensitive execution information of the agents. Consequently, the core challenge lies in enabling auditable task execution and fair incentive distribution for autonomous LLM agents in trustless environments, while simultaneously preserving their strategic privacy and minimizing on-chain costs. To address this challenge, we propose DAO-Agent, a novel framework that integrates three key technical innovations: (1) an on-chain decentralized autonomous organization (DAO) governance mechanism for transparent coordination and immutable logging; (2) a ZKP mechanism approach that enables Shapley-based contribution measurement off-chain, and (3) a hybrid on-chain/off-chain architecture that verifies ZKP-validated contribution measurements on-chain with minimal computational overhead. We implement DAO-Agent and conduct end-to-end experiments using a crypto trading task as a case study. Experimental results demonstrate that DAO-Agent achieves up to 99.9% reduction in verification gas costs compared to naive on-chain alternatives, with constant-time verification complexity that remains stable as coalition size increases, thereby establishing a scalable foundation for agent coordination in decentralized environments.

Method: AgentReuse uses intent classification to evaluate semantic similarities between requests, enabling plan reuse. It analyzes similarities and differences in request semantics rather than directly comparing original request texts.

Result: Achieves 93% effective plan reuse rate, F1 score of 0.9718, accuracy of 0.9459 in request similarity evaluation, and reduces latency by 93.12% compared to baselines without reuse mechanism.

Conclusion: AgentReuse effectively addresses the latency problem in LLM-driven agents by enabling plan reuse through semantic similarity analysis, significantly improving user experience while maintaining high accuracy in request matching.

Abstract: Integrating large language models (LLMs) into personal assistants, like Xiao Ai and Blue Heart V, effectively enhances their ability to interact with humans, solve complex tasks, and manage IoT devices. Such assistants are also termed LLM-driven agents. Upon receiving user requests, the LLM-driven agent generates plans using an LLM, executes these plans through various tools, and then returns the response to the user. During this process, the latency for generating a plan with an LLM can reach tens of seconds, significantly degrading user experience. Real-world dataset analysis shows that about 30% of the requests received by LLM-driven agents are identical or similar, which allows the reuse of previously generated plans to reduce latency. However, it is difficult to accurately define the similarity between the request texts received by the LLM-driven agent through directly evaluating the original request texts. Moreover, the diverse expressions of natural language and the unstructured format of plan texts make implementing plan reuse challenging. To address these issues, we present and implement a plan reuse mechanism for LLM-driven agents called AgentReuse. AgentReuse leverages the similarities and differences among requests’ semantics and uses intent classification to evaluate the similarities between requests and enable the reuse of plans. Experimental results based on a real-world dataset demonstrate that AgentReuse achieves a 93% effective plan reuse rate, an F1 score of 0.9718, and an accuracy of 0.9459 in evaluating request similarities, reducing latency by 93.12% compared with baselines without using the reuse mechanism.

Method: The paper develops computational foundations that formalize two critical dimensions of coopetition: (1) interdependence grounded in i* structural dependency analysis, translating depender-dependee-dependum relationships into quantitative interdependence coefficients via a structured translation framework; (2) complementarity formalized following Brandenburger and Nalebuff’s Added Value concept, modeling synergistic value creation with validated parameterization. The approach integrates structural dependencies with bargaining power in value appropriation and introduces a game-theoretic formulation where Nash Equilibrium incorporates structural interdependence.

Result: Validation involved over 22,000 experimental trials across power and logarithmic specifications using the Samsung-Sony S-LCD joint venture (2004-2011). Under strict historical alignment scoring, logarithmic specifications achieved 58/60 compared to power functions (46/60), producing realistic 41% cooperation increases aligning with documented S-LCD patterns, while power functions produced 166% increases exceeding realistic bounds. Statistical significance was confirmed at p < 0.001 with Cohen’s d > 9.

Conclusion: The paper successfully bridges the gap between qualitative strategic modeling and quantitative game theory for coopetition analysis. The computational foundations provide a rigorous framework for analyzing dynamic trade-offs in strategic coopetition, with logarithmic specifications demonstrating superior performance in capturing realistic coopetition patterns compared to power functions.

Abstract: Coopetition refers to simultaneous cooperation and competition among actors wherein actors ‘cooperate to grow the pie and compete to split it up.’ Modern socio-technical systems are characterized by strategic coopetition wherein actors concomitantly cooperate to create value and compete to capture it. While conceptual modeling languages such as i* provide rich qualitative representations of strategic dependencies, they lack mechanisms for quantitative analysis of dynamic trade-offs. Conversely, classical game theory offers mathematical rigor but strips away contextual richness. This report bridges this gap by developing computational foundations that formalize two critical dimensions of coopetition: interdependence and complementarity. We ground interdependence in i* structural dependency analysis, translating depender-dependee-dependum relationships into quantitative interdependence coefficients via a structured translation framework. We formalize complementarity following Brandenburger and Nalebuff’s Added Value concept, modeling synergistic value creation with validated parameterization. We integrate structural dependencies with bargaining power in value appropriation and introduce a game-theoretic formulation where Nash Equilibrium incorporates structural interdependence. Validation combines over 22,000 experimental trials across power and logarithmic specifications with the Samsung-Sony S-LCD joint venture (2004-2011). Under strict historical alignment scoring, logarithmic specifications achieve 58/60 compared to power functions (46/60), producing realistic 41% cooperation increases aligning with documented S-LCD patterns while power functions produce 166% increases exceeding realistic bounds. Statistical significance confirmed at p < 0.001, Cohen’s d > 9.

Method: Two-stage decoder-only generative LM: Stage-1 predicts coarse semantic tokens, Stage-2 generates fine acoustic tokens. Uses continuous SSL/codec embeddings, Frozen-LM Conditioning to reduce exposure bias, and DPO for human perceptual alignment.

Result: Experiments on Libri2Mix show GenTSE surpasses previous LM-based systems in speech quality, intelligibility, and speaker consistency.

Conclusion: Separating semantics and acoustics in a two-stage generative approach stabilizes decoding and yields more faithful, content-aligned target speech, advancing LM-based TSE performance.

Abstract: Language Model (LM)-based generative modeling has emerged as a promising direction for TSE, offering potential for improved generalization and high-fidelity speech. We present GenTSE, a two-stage decoder-only generative LM approach for TSE: Stage-1 predicts coarse semantic tokens, and Stage-2 generates fine acoustic tokens. Separating semantics and acoustics stabilizes decoding and yields more faithful, content-aligned target speech. Both stages use continuous SSL or codec embeddings, offering richer context than discretized-prompt methods. To reduce exposure bias, we employ a Frozen-LM Conditioning training strategy that conditions the LMs on predicted tokens from earlier checkpoints to reduce the gap between teacher-forcing training and autoregressive inference. We further employ DPO to better align outputs with human perceptual preferences. Experiments on Libri2Mix show that GenTSE surpasses previous LM-based systems in speech quality, intelligibility, and speaker consistency.

Method: Two complementary components: 1) Encoder-Decoder Attractor (EDA) network that automatically infers source count and acoustic clues for sound separation, and 2) multi-modal fusion network that interprets diverse user-provided clues (acoustic, semantic, or visual) for target sound extraction. Joint training with cross-task consistency constraints establishes a unified latent space.

Result: Remarkable performance in both tasks: 1.4 dB SDR improvement in sound separation compared to baseline and 86% target sound extraction accuracy.

Conclusion: The proposed unified framework successfully bridges sound separation and target sound extraction paradigms, enabling adaptive operation in either fully autonomous SS mode or clue-driven TSE mode with significant performance improvements.

Abstract: Sound separation (SS) and target sound extraction (TSE) are fundamental techniques for addressing complex acoustic scenarios. While existing SS methods struggle with determining the unknown number of sound sources, TSE approaches require precisely specified clues to achieve optimal performance. This paper proposes a unified framework that synergistically combines SS and TSE to overcome their individual limitations. Our architecture employs two complementary components: 1) An Encoder-Decoder Attractor (EDA) network that automatically infers both the source count and corresponding acoustic clues for SS, and 2) A multi-modal fusion network that precisely interprets diverse user-provided clues (acoustic, semantic, or visual) for TSE. Through joint training with cross-task consistency constraints, we establish a unified latent space that bridges both paradigms. During inference, the system adaptively operates in either fully autonomous SS mode or clue-driven TSE mode. Experiments demonstrate remarkable performance in both tasks, with notable improvements of 1.4 dB SDR improvement in SS compared to baseline and 86% TSE accuracy.

Method: Enhanced ASCHOPLEX integrated into Dafne (Deep Anatomical Federated Network) framework for federated incremental learning across 5 independent MRI datasets (2,284 subjects including MS patients and healthy controls).

Result: Federated incremental learning consistently achieves higher generalizability and more stable performance across diverse acquisition settings compared to conventional fine-tuning, which only works well on homogeneous data.

Conclusion: Federated incremental learning provides a robust alternative to conventional fine-tuning for choroid plexus segmentation, enabling better adaptation to heterogeneous MRI data from multiple sources and sequences.

Abstract: The Choroid Plexus (ChP) is a highly vascularized brain structure that plays a critical role in several physiological processes. ASCHOPLEX, a deep learning-based segmentation toolbox with an integrated fine-tuning stage, provides accurate ChP delineations on non-contrast-enhanced T1-weighted MRI scans; however, its performance is hindered by inter-dataset variability. This study introduces the first federated incremental learning approach for automated ChP segmentation from 3D T1-weighted brain MRI, by integrating an enhanced version of ASCHOPLEX within the Dafne (Deep Anatomical Federated Network) framework. A comparative evaluation is conducted to assess whether federated incremental learning through Dafne improves model generalizability across heterogeneous imaging conditions, relative to the conventional fine-tuning strategy employed by standalone ASCHOPLEX. The experimental cohort comprises 2,284 subjects, including individuals with Multiple Sclerosis as well as healthy controls, collected from five independent MRI datasets. Results indicate that the fine-tuning strategy provides high performance on homogeneous data (e.g., same MRI sequence, same cohort of subjects), but limited generalizability when the data variability is high (e.g., multiple MRI sequences, multiple and new cohorts of subjects). By contrast, the federated incremental learning variant of ASCHOPLEX constitutes a robust alternative consistently achieving higher generalizability and more stable performance across diverse acquisition settings.

Method: Instance-specific optimization that directly optimizes channel symbols and a lightweight neural decoder for each source, using overfitting to eliminate the need for training data or pre-trained models.

Result: Achieves around 1000x lower decoding complexity than alternatives, using only 607 model parameters and 641 multiplications per pixel, with state-of-the-art performance in high SNR regimes.

Conclusion: Implicit-JSCC shows promise for future communication systems, especially streaming scenarios where one-time offline encoding supports multiple online decoding, offering an efficient, modality-agnostic solution.

Abstract: This paper introduces Implicit-JSCC, a novel overfitted joint source-channel coding paradigm that directly optimizes channel symbols and a lightweight neural decoder for each source. This instance-specific strategy eliminates the need for training datasets or pre-trained models, enabling a storage-free, modality-agnostic solution. As a low-complexity alternative, Implicit-JSCC achieves efficient image transmission with around 1000x lower decoding complexity, using as few as 607 model parameters and 641 multiplications per pixel. This overfitted design inherently addresses source generalizability and achieves state-of-the-art results in the high SNR regimes, underscoring its promise for future communication systems, especially streaming scenarios where one-time offline encoding supports multiple online decoding.

Method: Proposed SL-Net, a single-shot deep learning model that removes solar loading transients from thermal facial images. Uses co-registered RGB-thermal images from a diverse dataset of 100 subjects with IRT and skin tone measurements.

Result: Forehead skin temperature increases by 2.00°C after solar loading. SL-Net reduces this error by 68% to 0.64°C. The model eliminates skin tone bias in IRT performance that exists with standard solar loading effects.

Conclusion: Machine learning can correct complex thermal perturbations for robust and equitable human thermography. The work demonstrates feasibility of single-shot correction for solar loading effects, addressing both accuracy and fairness issues in IRT fever detection.

Abstract: Widely deployed for fever detection, infrared thermometers (IRTs) enable rapid non-contact measurement of core body temperature but are inaccurate in unconstrained environments when skin temperature is transient. In this work, we present the first study on the effect of solar loading–solar radiation-induced elevation of skin but not core temperature–on IRT performance. Solar loading causes poor specificity in IRT fever detection, and the standard procedure is to reacclimate subjects for up to 30 minutes before IRT measurement. In contrast, we propose a single-shot deep learning model that removes solar loading transients from thermal facial images, allowing accurate IRT operation in solar loaded conditions. Forehead skin temperature increases by 2.00°C after solar loading, and our deep learning model, SL-Net, reduces this error by 68% to 0.64°C. We show that the solar loading effect depends on skin tone, introducing inequity in IRT performance, while SL-Net is unbiased. We open source a diverse dataset of 100 subjects with co-registered RGB-thermal images, and IRT and skin tone measurements. Our work shows that it is possible to use machine learning to correct complex thermal perturbations to enable robust and equitable human thermography.

Method: V-Rex introduces ReSV, a training-free dynamic KV cache retrieval algorithm that exploits temporal and spatial similarity-based token clustering to reduce KV cache memory across frames. It also provides a compact hardware accelerator with a dynamic KV cache retrieval engine (DRE) featuring bit-level and early-exit computing units.

Result: Achieves 3.9-8.3 FPS real-time performance on edge deployment with negligible accuracy loss. DRE accounts for only 2.2% power and 2.0% area while delivering 1.9-19.7x speedup and 3.1-18.5x energy efficiency improvements over AGX Orin GPU.

Conclusion: V-Rex is the first comprehensive solution tackling KV cache retrieval across both algorithms and hardware, enabling real-time streaming video LLM inference on resource-constrained edge devices.

Abstract: Streaming video large language models (LLMs) are increasingly used for real-time multimodal tasks such as video captioning, question answering, conversational agents, and augmented reality. However, these models face fundamental memory and computational challenges because their key-value (KV) caches grow substantially with continuous streaming video input. This process requires an iterative prefill stage, which is a unique feature of streaming video LLMs. Due to its iterative prefill stage, it suffers from significant limitations, including extensive computation, substantial data transfer, and degradation in accuracy. Crucially, this issue is exacerbated for edge deployment, which is the primary target for these models. In this work, we propose V-Rex, the first software-hardware co-designed accelerator that comprehensively addresses both algorithmic and hardware bottlenecks in streaming video LLM inference. At its core, V-Rex introduces ReSV, a training-free dynamic KV cache retrieval algorithm. ReSV exploits temporal and spatial similarity-based token clustering to reduce excessive KV cache memory across video frames. To fully realize these algorithmic benefits, V-Rex offers a compact, low-latency hardware accelerator with a dynamic KV cache retrieval engine (DRE), featuring bit-level and early-exit based computing units. V-Rex achieves unprecedented real-time of 3.9-8.3 FPS and energy-efficient streaming video LLM inference on edge deployment with negligible accuracy loss. While DRE only accounts for 2.2% power and 2.0% area, the system delivers 1.9-19.7x speedup and 3.1-18.5x energy efficiency improvements over AGX Orin GPU. This work is the first to comprehensively tackle KV cache retrieval across algorithms and hardware, enabling real-time streaming video LLM inference on resource-constrained edge devices.

Method: Created a multi-center breast MRI dataset from 6 clinical institutions across 5 European countries, comprising 741 examinations with malignant, benign, and non-lesion cases. Data includes heterogeneous scanners, field strengths, and protocols reflecting real-world variability.

Result: Presented a publicly available dataset and conducted baseline benchmark experiments using a transformer-based model to demonstrate potential use cases and provide reference performance for future comparisons.

Conclusion: This dataset addresses a critical gap in breast MRI AI research by providing diverse, real-world data that can accelerate development of AI tools for breast cancer detection and improve clinical scalability.

Abstract: Early detection of breast cancer is critical for improving patient outcomes. While mammography remains the primary screening modality, magnetic resonance imaging (MRI) is increasingly recommended as a supplemental tool for women with dense breast tissue and those at elevated risk. However, the acquisition and interpretation of multiparametric breast MRI are time-consuming and require specialized expertise, limiting scalability in clinical practice. Artificial intelligence (AI) methods have shown promise in supporting breast MRI interpretation, but their development is hindered by the limited availability of large, diverse, and publicly accessible datasets. To address this gap, we present a publicly available, multi-center breast MRI dataset collected across six clinical institutions in five European countries. The dataset comprises 741 examinations from women undergoing screening or diagnostic breast MRI and includes malignant, benign, and non-lesion cases. Data were acquired using heterogeneous scanners, field strengths, and acquisition protocols, reflecting real-world clinical variability. In addition, we report baseline benchmark experiments using a transformer-based model to illustrate potential use cases of the dataset and to provide reference performance for future methodological comparisons.