# Research Papers — 2025-10 · bundle page 146 of 217 > Full-text markdown of papers #7251–7300 of 10843 published in 2025-10, newest first. AI-readable mirrors of the full 10843-paper 2025-10 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2025-10/page/146 [← Previous bundle](https://aipod.dev/papers/archive/2025-10/page/145.md) [Next bundle →](https://aipod.dev/papers/archive/2025-10/page/147.md) ### A Model-Driven Engineering Approach to AI-Powered Healthcare Platforms Authors: Mira Raheem, Amal Elgammal, Michael Papazoglou, Bernd Krämer, Neamat El-Tazi Published: 2025-10-10 Categories: cs.SE, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2510.09308v1 Artificial intelligence (AI) has the potential to transform healthcare by supporting more accurate diagnoses and personalized treatments. However, its adoption in practice remains constrained by fragmented data sources, strict privacy rules, and the technical complexity of building reliable clinical systems. To address these challenges, we introduce a model driven engineering (MDE) framework designed specifically for healthcare AI. The framework relies on formal metamodels, domain-specific languages (DSLs), and automated transformations to move from high level specifications to running software. At its core is the Medical Interoperability Language (MILA), a graphical DSL that enables clinicians and data scientists to define queries and machine learning pipelines using shared ontologies. When combined with a federated learning architecture, MILA allows institutions to collaborate without exchanging raw patient data, ensuring semantic consistency across sites while preserving privacy. We evaluate this approach in a multi center cancer immunotherapy study. The generated pipelines delivered strong predictive performance, with support vector machines achieving up to 98.5 percent and 98.3 percent accuracy in key tasks, while substantially reducing manual coding effort. These findings suggest that MDE principles metamodeling, semantic integration, and automated code generation can provide a practical path toward interoperable, reproducible, and trustworthy digital health platforms. --- ### Target speaker anonymization in multi-speaker recordings Authors: Natalia Tomashenko, Junichi Yamagishi, Xin Wang, Yun Liu, Emmanuel Vincent Published: 2025-10-10 Categories: eess.AS, cs.CL, cs.CR arXiv: https://arxiv.org/abs/2510.09307v1 Most of the existing speaker anonymization research has focused on single-speaker audio, leading to the development of techniques and evaluation metrics optimized for such condition. This study addresses the significant challenge of speaker anonymization within multi-speaker conversational audio, specifically when only a single target speaker needs to be anonymized. This scenario is highly relevant in contexts like call centers, where customer privacy necessitates anonymizing only the customer's voice in interactions with operators. Conventional anonymization methods are often not suitable for this task. Moreover, current evaluation methodology does not allow us to accurately assess privacy protection and utility in this complex multi-speaker scenario. This work aims to bridge these gaps by exploring effective strategies for targeted speaker anonymization in conversational audio, highlighting potential problems in their development and proposing corresponding improved evaluation methodologies. --- ### Rewiring Development in Brain Segmentation: Leveraging Adult Brain Priors for Enhancing Infant MRI Segmentation Authors: Alemu Sisay Nigru, Michele Svanera, Austin Dibble, Connor Dalby, Mattia Savardi, Sergio Benini Published: 2025-10-10 Categories: eess.IV, cs.CV arXiv: https://arxiv.org/abs/2510.09306v1 Accurate segmentation of infant brain MRI is critical for studying early neurodevelopment and diagnosing neurological disorders. Yet, it remains a fundamental challenge due to continuously evolving anatomy of the subjects, motion artifacts, and the scarcity of high-quality labeled data. In this work, we present LODi, a novel framework that utilizes prior knowledge from an adult brain MRI segmentation model to enhance the segmentation performance of infant scans. Given the abundance of publicly available adult brain MRI data, we pre-train a segmentation model on a large adult dataset as a starting point. Through transfer learning and domain adaptation strategies, we progressively adapt the model to the 0-2 year-old population, enabling it to account for the anatomical and imaging variability typical of infant scans. The adaptation of the adult model is carried out using weakly supervised learning on infant brain scans, leveraging silver-standard ground truth labels obtained with FreeSurfer. By introducing a novel training strategy that integrates hierarchical feature refinement and multi-level consistency constraints, our method enables fast, accurate, age-adaptive segmentation, while mitigating scanner and site-specific biases. Extensive experiments on both internal and external datasets demonstrate the superiority of our approach over traditional supervised learning and domain-specific models. Our findings highlight the advantage of leveraging adult brain priors as a foundation for age-flexible neuroimaging analysis, paving the way for more reliable and generalizable brain MRI segmentation across the lifespan. --- ### CapGeo-Bench: Decoupling Visual Perception from Reasoning and Evaluating Geometric Understanding Authors: Yuying Li, Siyi Qian, Hao Liang, Leqi Zheng, Ruichuan An, Linzhuang Sun, Jiajun Zhang, Wentao Zhang Published: 2025-10-10 Categories: cs.CV, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2510.09302v2 While Multimodal Large Language Models (MLLMs) have achieved remarkable success in difficult purely textual mathematical reasoning tasks, even advanced closed-source models such as GPT-o3 still struggle with geometric problems. This discrepancy motivates us to investigate the root cause: is the bottleneck of multimodal geometric reasoning rooted in reasoning itself, or in the perception of geometric information from diagrams? To answer this question, we conduct an exploratory experiment and find that providing high-quality captions consistently and substantially boosts performance of MLLMs, empirically validating the visual perception bottleneck in geometric reasoning. However, MLLMs' capabilities in visual geometric perception remain insufficiently evaluated. To fill the gap, we propose CapGeo-Bench, a benchmark of 4,641 high-quality figure-caption pairs equipped with a fine-grained keypoint-based evaluation metric. Specifically, CapGeo-Bench encompasses three geometric classes and is divided into four difficulty levels according to the complexity of geometric primitives. Captions are scored at a fine-grained level by extracting keypoints along three dimensions-geometric elements, spatial relationships, and numerical attributes. CapGeo-Bench provides a reliable instrument and guideline for evaluating and advancing MLLMs' geometric perception capabilities. Our code and data are publicly available at: https://github.com/YuYingLi0/CapGeo. --- ### Foraging with the Eyes: Dynamics in Human Visual Gaze and Deep Predictive Modeling Authors: Tejaswi V. Panchagnula Published: 2025-10-10 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2510.09299v1 Animals often forage via Levy walks stochastic trajectories with heavy tailed step lengths optimized for sparse resource environments. We show that human visual gaze follows similar dynamics when scanning images. While traditional models emphasize image based saliency, the underlying spatiotemporal statistics of eye movements remain underexplored. Understanding these dynamics has broad applications in attention modeling and vision-based interfaces. In this study, we conducted a large scale human subject experiment involving 40 participants viewing 50 diverse images under unconstrained conditions, recording over 4 million gaze points using a high speed eye tracker. Analysis of these data shows that the gaze trajectory of the human eye also follows a Levy walk akin to animal foraging. This suggests that the human eye forages for visual information in an optimally efficient manner. Further, we trained a convolutional neural network (CNN) to predict fixation heatmaps from image input alone. The model accurately reproduced salient fixation regions across novel images, demonstrating that key components of gaze behavior are learnable from visual structure alone. Our findings present new evidence that human visual exploration obeys statistical laws analogous to natural foraging and open avenues for modeling gaze through generative and predictive frameworks. --- ### ShiZhi: A Chinese Lightweight Large Language Model for Court View Generation Authors: Zhitian Hou, Kun Zeng Published: 2025-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.09297v2 Criminal Court View Generation (CVG) is a fundamental task in legal artificial intelligence, aiming to automatically generate the "Court View" section of a legal case document. Generating court views is challenging due to the diversity and complexity of case facts, and directly generating from raw facts may limit performance. In this paper, we present ShiZhi, the first large language model (LLM) specifically designed for court view generation. We construct a Chinese Court View Generation dataset, CCVG, of more than 110K cases, each containing fact descriptions paired with corresponding court views. Based on this dataset, ShiZhi achieving 70.00 ROUGE-1 and 67.85 BLEU-1 on court view generation, as well as 86.48\% accuracy with 92.75\% macro F1 on charge prediction. Experimental results demonstrate that even a small LLM can generate reasonable and legally coherent court views when trained on high-quality domain-specific data. Our model and dataset are available at \href{https://github.com/ZhitianHou/ShiZhi}{https://github.com/ZhitianHou/ShiZhi}. --- ### MaP: A Unified Framework for Reliable Evaluation of Pre-training Dynamics Authors: Jiapeng Wang, Changxin Tian, Kunlong Chen, Ziqi Liu, Jiaxin Mao, Wayne Xin Zhao, Zhiqiang Zhang, Jun Zhou Published: 2025-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.09295v2 Reliable evaluation is fundamental to the progress of Large Language Models (LLMs), yet the evaluation process during pre-training is plagued by significant instability that obscures true learning dynamics. In this work, we systematically diagnose this instability, attributing it to two distinct sources: \textit{Parameter Instability} from training stochasticity and \textit{Evaluation Instability} from noisy measurement protocols. To counteract both sources of noise, we introduce \textbf{MaP}, a dual-pronged framework that synergistically integrates checkpoint \underline{M}erging \underline{a}nd the \underline{P}ass@k metric. Checkpoint merging smooths the parameter space by averaging recent model weights, while Pass@k provides a robust, low-variance statistical estimate of model capability. Extensive experiments show that MaP yields significantly smoother performance curves, reduces inter-run variance, and ensures more consistent model rankings. Ultimately, MaP provides a more reliable and faithful lens for observing LLM training dynamics, laying a crucial empirical foundation for LLM research. --- ### Mitigating Model Drift in Developing Economies Using Synthetic Data and Outliers Authors: Ilyas Varshavskiy, Bonu Boboeva, Shuhrat Khalilbekov, Azizjon Azimi, Sergey Shulgin, Akhlitdin Nizamitdinov, Haitz Sáez de Ocáriz Borde Published: 2025-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.09294v2 Machine Learning models in finance are highly susceptible to model drift, where predictive performance declines as data distributions shift. This issue is especially acute in developing economies such as those in Central Asia and the Caucasus - including Tajikistan, Uzbekistan, Kazakhstan, and Azerbaijan - where frequent and unpredictable macroeconomics shocks destabilize financial data. To the best of our knowledge, this is among the first studies to examine drift mitigation methods on financial datasets from these regions. We investigate the use of synthetic outliers, a largely unexplored approach, to improve model stability against unforeseen shocks. To evaluate effectiveness, we introduce a two-level framework that measures both the extent of performance degradation and the severity of shocks. Our experiments on macroeconomic tabular datasets show that adding a small proportion of synthetic outliers generally improves stability compared to baseline models, though the optimal amount varies by dataset and model --- ### One Sentence, Two Embeddings: Contrastive Learning of Explicit and Implicit Semantic Representations Authors: Kohei Oda, Po-Min Chuang, Kiyoaki Shirai, Natthawut Kertkeidkachorn Published: 2025-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.09293v2 Sentence embedding methods have made remarkable progress, yet they still struggle to capture the implicit semantics within sentences. This can be attributed to the inherent limitations of conventional sentence embedding methods that assign only a single vector per sentence. To overcome this limitation, we propose DualCSE, a sentence embedding method that assigns two embeddings to each sentence: one representing the explicit semantics and the other representing the implicit semantics. These embeddings coexist in the shared space, enabling the selection of the desired semantics for specific purposes such as information retrieval and text classification. Experimental results demonstrate that DualCSE can effectively encode both explicit and implicit meanings and improve the performance of the downstream task. --- ### A unifying Bayesian framework for adversarial robustness Authors: Pablo G. Arce, Roi Naveiro, David Ríos Insua Published: 2025-10-10 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2510.09288v2 The vulnerability of machine learning models to adversarial attacks remains a critical societal security challenge. Traditional defenses, such as adversarial training, typically robustify models by minimizing a worst-case loss. These deterministic approaches do not account for uncertainty in the adversary's attack. While stochastic defenses placing a probability distribution on the adversary exist, they often lack statistical rigor and fail to make explicit their underlying assumptions. To resolve these issues, we introduce a formal Bayesian framework that models adversarial uncertainty through a stochastic channel, articulating all probabilistic assumptions. This yields two robustification strategies: a proactive defense enacted during training, aligned with adversarial training, and a reactive defense enacted during operations, aligned with adversarial purification. Several state-of-the-art defenses can be recovered as limiting cases of our model. We empirically validate our methodology, showcasing the benefits of explicitly modeling adversarial uncertainty. --- ### Spotlight on Token Perception for Multimodal Reinforcement Learning Authors: Siyuan Huang, Xiaoye Qu, Yafu Li, Yun Luo, Zefeng He, Daizong Liu, Yu Cheng Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09285v2 While Reinforcement Learning with Verifiable Rewards (RLVR) has advanced the reasoning capabilities of Large Vision-Language Models (LVLMs), most existing methods in multimodal reasoning neglect the critical role of visual perception within the RLVR optimization process. In this paper, we undertake a pioneering exploration of multimodal RLVR through the novel perspective of token perception, which measures the visual dependency of each generated token. With a granular analysis of Chain-of-Thought (CoT) processes, we uncover two key insights: first, token perception in a rollout trajectory is sparsely distributed, where only a small fraction of tokens have high visual dependency for visually-grounded reasoning; second, different trajectories exhibit significant divergence in their overall visual dependency. Based on these observations, we propose Visually-Perceptive Policy Optimization (VPPO), a novel policy gradient algorithm that explicitly leverages token perception to refine the learning signal. Specifically, VPPO achieves this through a dual mechanism: it reweights a trajectory's advantage by its overall visual dependency, and focuses policy updates exclusively on perceptually pivotal tokens. On a comprehensive suite of eight perception and reasoning benchmarks, VPPO demonstrates substantial gains over leading open-source RL-tuned models, with its effectiveness consistently validated across 7B and 32B model scales. Our findings not only establish a new token-level perceptual perspective for analyzing multimodal RLVR but also present a novel and effective optimization strategy to significantly enhance the multimodal reasoning capabilities of LVLMs. --- ### CLARity: Reasoning Consistency Alone Can Teach Reinforced Experts Authors: Jiuheng Lin, Cong Jiang, Zirui Wu, Jiarui Sun, Yansong Feng Published: 2025-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.09278v2 Training expert LLMs in domains with scarce data is difficult, often relying on multiple-choice questions (MCQs). However, standard outcome-based reinforcement learning (RL) on MCQs is risky. While it may improve accuracy, we observe it often degrades reasoning quality such as logical consistency. Existing solutions to supervise reasoning, such as large-scale Process Reward Models (PRMs), are prohibitively expensive. To address this, we propose CLARity, a cost-effective RL framework that enhances reasoning quality using only a small, general-purpose LLM. CLARity integrates a consistency-aware reward mechanism with a 2-stage refine-then-monitor training pipeline to enhance reasoning consistency, and a dynamic data reformulation strategy to to better exploit limited data. Experiments demonstrate that CLARity improves response consistency by 16.5% and accuracy by 7.5% over baselines. Human evaluations further confirm holistic improvements in coherence and professionalism. Thus, CLARity offers a generalizable solution that enables smaller models to effectively guide expert models by reasoning consistency. Our code is open sourced at: https://github.com/Infinite-set/CLARity --- ### Inflated Excellence or True Performance? Rethinking Medical Diagnostic Benchmarks with Dynamic Evaluation Authors: Xiangxu Zhang, Lei Li, Yanyun Zhou, Xiao Zhou, Yingying Zhang, Xian Wu Published: 2025-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.09275v2 Medical diagnostics is a high-stakes and complex domain that is critical to patient care. However, current evaluations of large language models (LLMs) remain limited in capturing key challenges of clinical diagnostic scenarios. Most rely on benchmarks derived from public exams, raising contamination bias that can inflate performance, and they overlook the confounded nature of real consultations beyond textbook cases. Recent dynamic evaluations offer a promising alternative, but often remain insufficient for diagnosis-oriented benchmarking, with limited coverage of clinically grounded confounders and trustworthiness beyond accuracy. To address these gaps, we propose DyReMe, a dynamic benchmark for medical diagnostics that provides a controlled and scalable stress test of diagnostic robustness. Unlike static exam-style questions, DyReMe generates fresh, consultation-style cases that incorporate clinically grounded confounders, such as differential diagnoses and common misdiagnosis factors. It also varies expression styles to capture heterogeneous patient-style descriptions. Beyond accuracy, DyReMe evaluates LLMs on three additional clinically relevant dimensions: veracity, helpfulness, and consistency. Our experiments show that this dynamic approach yields more challenging assessments and exposes substantial weaknesses of stateof-the-art LLMs under clinically confounded diagnostic settings. These findings highlight the urgent need for evaluation frameworks that better assess trustworthy medical diagnostics 1 under clinically grounded confounders. --- ### MomentSeg: Moment-Centric Sampling for Enhanced Video Pixel Understanding Authors: Ming Dai, Sen Yang, Boqiang Duan, Wankou Yang, Jingdong Wang Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09274v1 Referring Video Object Segmentation (RefVOS) seeks to segment target objects in videos guided by natural language descriptions, demanding both temporal reasoning and fine-grained visual comprehension. Existing sampling strategies for LLM-based approaches typically rely on either handcrafted heuristics or external keyframe models. The former often overlooks essential temporal cues, while the latter increases system complexity. To address this, we propose a unified framework that jointly optimizes Temporal Sentence Grounding (TSG) and RefVOS, naturally incorporating key moment grounding capability. During training, we introduce a novel TSG paradigm that employs a dedicated \texttt{(FIND)} token for key moment identification through temporal token similarity matching, thereby avoiding the need for external timestamp encodings. For inference, we design a Moment-Centric Sampling (MCS) strategy that densely samples informative moments while sparsely sampling non-essential frames, preserving both motion details and global context. To further enhance tracking stability, we develop Bidirectional Anchor-updated Propagation (BAP), which leverages the most relevant moment as start point for high-quality mask initialization and dynamically updates at sampled points to mitigate accumulated errors. Code and model will be available at: https://github.com/Dmmm1997/MomentSeg --- ### Goal-oriented Backdoor Attack against Vision-Language-Action Models via Physical Objects Authors: Zirun Zhou, Zhengyang Xiao, Haochuan Xu, Jing Sun, Di Wang, Jingfeng Zhang Published: 2025-10-10 Categories: cs.CR, cs.CV, cs.LG arXiv: https://arxiv.org/abs/2510.09269v1 Recent advances in vision-language-action (VLA) models have greatly improved embodied AI, enabling robots to follow natural language instructions and perform diverse tasks. However, their reliance on uncurated training datasets raises serious security concerns. Existing backdoor attacks on VLAs mostly assume white-box access and result in task failures instead of enforcing specific actions. In this work, we reveal a more practical threat: attackers can manipulate VLAs by simply injecting physical objects as triggers into the training dataset. We propose goal-oriented backdoor attacks (GoBA), where the VLA behaves normally in the absence of physical triggers but executes predefined and goal-oriented actions in the presence of physical triggers. Specifically, based on a popular VLA benchmark LIBERO, we introduce BadLIBERO that incorporates diverse physical triggers and goal-oriented backdoor actions. In addition, we propose a three-level evaluation that categorizes the victim VLA's actions under GoBA into three states: nothing to do, try to do, and success to do. Experiments show that GoBA enables the victim VLA to successfully achieve the backdoor goal in 97 percentage of inputs when the physical trigger is present, while causing zero performance degradation on clean inputs. Finally, by investigating factors related to GoBA, we find that the action trajectory and trigger color significantly influence attack performance, while trigger size has surprisingly little effect. The code and BadLIBERO dataset are accessible via the project page at https://goba-attack.github.io/. --- ### Placeit! A Framework for Learning Robot Object Placement Skills Authors: Amina Ferrad, Johann Huber, François Hélénon, Julien Gleyze, Mahdi Khoramshahi, Stéphane Doncieux Published: 2025-10-10 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2510.09267v1 Robotics research has made significant strides in learning, yet mastering basic skills like object placement remains a fundamental challenge. A key bottleneck is the acquisition of large-scale, high-quality data, which is often a manual and laborious process. Inspired by Graspit!, a foundational work that used simulation to automatically generate dexterous grasp poses, we introduce Placeit!, an evolutionary-computation framework for generating valid placement positions for rigid objects. Placeit! is highly versatile, supporting tasks from placing objects on tables to stacking and inserting them. Our experiments show that by leveraging quality-diversity optimization, Placeit! significantly outperforms state-of-the-art methods across all scenarios for generating diverse valid poses. A pick&place pipeline built on our framework achieved a 90% success rate over 120 real-world deployments. This work positions Placeit! as a powerful tool for open-environment pick-and-place tasks and as a valuable engine for generating the data needed to train simulation-based foundation models in robotics. --- ### CFVBench: A Comprehensive Video Benchmark for Fine-grained Multimodal Retrieval-Augmented Generation Authors: Kaiwen Wei, Xiao Liu, Jie Zhang, Zijian Wang, Ruida Liu, Yuming Yang, Xin Xiao, Xiao Sun, Haoyang Zeng, Changzai Pan, Yidan Zhang, Jiang Zhong et al. Published: 2025-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.09266v1 Multimodal Retrieval-Augmented Generation (MRAG) enables Multimodal Large Language Models (MLLMs) to generate responses with external multimodal evidence, and numerous video-based MRAG benchmarks have been proposed to evaluate model capabilities across retrieval and generation stages. However, existing benchmarks remain limited in modality coverage and format diversity, often focusing on single- or limited-modality tasks, or coarse-grained scene understanding. To address these gaps, we introduce CFVBench, a large-scale, manually verified benchmark constructed from 599 publicly available videos, yielding 5,360 open-ended QA pairs. CFVBench spans high-density formats and domains such as chart-heavy reports, news broadcasts, and software tutorials, requiring models to retrieve and reason over long temporal video spans while maintaining fine-grained multimodal information. Using CFVBench, we systematically evaluate 7 retrieval methods and 14 widely-used MLLMs, revealing a critical bottleneck: current models (even GPT5 or Gemini) struggle to capture transient yet essential fine-grained multimodal details. To mitigate this, we propose Adaptive Visual Refinement (AVR), a simple yet effective framework that adaptively increases frame sampling density and selectively invokes external tools when necessary. Experiments show that AVR consistently enhances fine-grained multimodal comprehension and improves performance across all evaluated MLLMs --- ### SynthID-Image: Image watermarking at internet scale Authors: Sven Gowal, Rudy Bunel, Florian Stimberg, David Stutz, Guillermo Ortiz-Jimenez, Christina Kouridi, Mel Vecerik, Jamie Hayes, Sylvestre-Alvise Rebuffi, Paul Bernard, Chris Gamble, Miklós Z. Horváth et al. Published: 2025-10-10 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2510.09263v1 We introduce SynthID-Image, a deep learning-based system for invisibly watermarking AI-generated imagery. This paper documents the technical desiderata, threat models, and practical challenges of deploying such a system at internet scale, addressing key requirements of effectiveness, fidelity, robustness, and security. SynthID-Image has been used to watermark over ten billion images and video frames across Google's services and its corresponding verification service is available to trusted testers. For completeness, we present an experimental evaluation of an external model variant, SynthID-O, which is available through partnerships. We benchmark SynthID-O against other post-hoc watermarking methods from the literature, demonstrating state-of-the-art performance in both visual quality and robustness to common image perturbations. While this work centers on visual media, the conclusions on deployment, constraints, and threat modeling generalize to other modalities, including audio. This paper provides a comprehensive documentation for the large-scale deployment of deep learning-based media provenance systems. --- ### GREAT: Generalizable Backdoor Attacks in RLHF via Emotion-Aware Trigger Synthesis Authors: Subrat Kishore Dutta, Yuelin Xu, Piyush Pant, Xiao Zhang Published: 2025-10-10 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2510.09260v3 Recent work has shown that RLHF is highly susceptible to backdoor attacks. However, existing methods often rely on rare tokens or fixed triggers, limiting their impact in realistic scenarios. In this work, we develop GREAT, a novel framework for crafting natural distributional backdoors in RLHF. Specifically, GREAT targets harmful response generation for a vulnerable user subpopulation featured by semantically violent requests paired with emotionally angry triggers. At the core of our framework is a trigger identification pipeline that operates in the model's latent embedding space, leveraging dimensionality reduction and clustering techniques to identify representative triggers. To enable this, we introduce a hierarchical and diversity-driven prompting strategy to construct Erinyes, a high-quality dataset of over 5,000 angry triggers curated from GPT-4.1. Our experiments show that GREAT significantly outperforms baselines in attack generalization to unseen triggers, while preserving standard utility and maintaining stealth under defenses. --- ### Detecting Data Contamination from Reinforcement Learning Post-training for Large Language Models Authors: Yongding Tao, Tian Wang, Yihong Dong, Huanyu Liu, Kechi Zhang, Xiaolong Hu, Ge Li Published: 2025-10-10 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2510.09259v2 Data contamination poses a significant threat to the reliable evaluation of Large Language Models (LLMs). This issue arises when benchmark samples may inadvertently appear in training sets, compromising the validity of reported performance. While detection methods have been developed for the pre-training and Supervised Fine-Tuning stages, a critical research gap exists for the increasingly significant phase of Reinforcement Learning (RL) post-training. As RL post-training becomes pivotal for advancing LLM reasoning, the absence of specialized contamination detection methods in this paradigm presents a critical vulnerability. To address this, we conduct the first systematic study of data detection within RL post-training scenario and propose Self-Critique. Our method is motivated by a key observation: after RL phase, the output entropy distribution of LLMs tends to collapse into highly specific and sparse modes. Self-Critique probes for the underlying policy collapse, i.e., the model's convergence to a narrow reasoning path, which causes this entropy reduction. To facilitate this research, we also introduce RL-MIA, a benchmark constructed to simulate this specific contamination scenario. Extensive experiments show that Self-Critique significantly outperforms baseline methods across multiple models and contamination tasks, achieving an AUC improvement of up to 30%. Whereas existing methods are close to a random guess for RL-phase contamination, our method makes detection possible. --- ### Hallucination Filtering in Radiology Vision-Language Models Using Discrete Semantic Entropy Authors: Patrick Wienholt, Sophie Caselitz, Robert Siepmann, Philipp Bruners, Keno Bressem, Christiane Kuhl, Jakob Nikolas Kather, Sven Nebelung, Daniel Truhn Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09256v2 To determine whether using discrete semantic entropy (DSE) to reject questions likely to generate hallucinations can improve the accuracy of black-box vision-language models (VLMs) in radiologic image based visual question answering (VQA). This retrospective study evaluated DSE using two publicly available, de-identified datasets: the VQA-Med 2019 benchmark (500 images with clinical questions and short-text answers) and a diagnostic radiology dataset (206 cases: 60 computed tomography scans, 60 magnetic resonance images, 60 radiographs, 26 angiograms) with corresponding ground-truth diagnoses. GPT-4o and GPT-4.1 (Generative Pretrained Transformer; OpenAI) answered each question 15 times using a temperature of 1.0. Baseline accuracy was determined using low-temperature answers (temperature 0.1). Meaning-equivalent responses were grouped using bidirectional entailment checks, and DSE was computed from the relative frequencies of the resulting semantic clusters. Accuracy was recalculated after excluding questions with DSE > 0.6 or > 0.3. p-values and 95% confidence intervals were obtained using bootstrap resampling and a Bonferroni-corrected threshold of p < .004 for statistical significance. Across 706 image-question pairs, baseline accuracy was 51.7% for GPT-4o and 54.8% for GPT-4.1. After filtering out high-entropy questions (DSE > 0.3), accuracy on the remaining questions was 76.3% (retained questions: 334/706) for GPT-4o and 63.8% (retained questions: 499/706) for GPT-4.1 (both p < .001). Accuracy gains were observed across both datasets and largely remained statistically significant after Bonferroni correction. DSE enables reliable hallucination detection in black-box VLMs by quantifying semantic inconsistency. This method significantly improves diagnostic answer accuracy and offers a filtering strategy for clinical VLM applications. --- ### DSPO: Stable and Efficient Policy Optimization for Agentic Search and Reasoning Authors: Chenyang Gu, Yewen Pu, Bruce Yang, Xiaofan Li, Huan Gao Published: 2025-10-10 Categories: cs.CL arXiv: https://arxiv.org/abs/2510.09255v4 Enhancing LLMs with the ability to actively search external knowledge is crucial for complex and real-world tasks. Current approaches either rely on prompting to elicit the model's innate agent capabilities, or suffer from performance ceilings and collapse when applying RL to complex interactive tasks, leaving their true agentic potential untapped. To address this, we introduce \textbf{D}ynamic-filter \textbf{S}equence-level \textbf{P}olicy \textbf{O}ptimization (DSPO), an improved RL algorithm designed for robust agent training through sequence-level optimization and dynamic sample filtering. We train our model purely through RL to interleave multi-turn search and reasoning, obviating the need for supervised demonstration data. Across multiple QA benchmarks, our 7B model improves over a comparable previous work by \textbf{34.1\%}, and even outperforms the 14B model from previous work in complex multihop QA such as HotpotQA by nearly \textbf{9\% relative}, maintaining exceptional training stability. --- ### Obstacle Avoidance using Dynamic Movement Primitives and Reinforcement Learning Authors: Dominik Urbaniak, Alejandro Agostini, Pol Ramon, Jan Rosell, Raúl Suárez, Michael Suppa Published: 2025-10-10 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2510.09254v1 Learning-based motion planning can quickly generate near-optimal trajectories. However, it often requires either large training datasets or costly collection of human demonstrations. This work proposes an alternative approach that quickly generates smooth, near-optimal collision-free 3D Cartesian trajectories from a single artificial demonstration. The demonstration is encoded as a Dynamic Movement Primitive (DMP) and iteratively reshaped using policy-based reinforcement learning to create a diverse trajectory dataset for varying obstacle configurations. This dataset is used to train a neural network that takes as inputs the task parameters describing the obstacle dimensions and location, derived automatically from a point cloud, and outputs the DMP parameters that generate the trajectory. The approach is validated in simulation and real-robot experiments, outperforming a RRT-Connect baseline in terms of computation and execution time, as well as trajectory length, while supporting multi-modal trajectory generation for different obstacle geometries and end-effector dimensions. Videos and the implementation code are available at https://github.com/DominikUrbaniak/obst-avoid-dmp-pi2. --- ### On the Robustness of Vision-Language Models in Zero-shot Privacy Classification Authors: Alina Elena Baia, Alessio Xompero, Andrea Cavallaro Published: 2025-10-10 Categories: cs.CV, cs.LG, cs.MM arXiv: https://arxiv.org/abs/2510.09253v2 Automatic systems for document understanding require multimodal models that accurately identify sensitive visual content, even in the presence of image degradations. Instruction-following large Vision-Language Models (VLMs) are expected to generalise across domains and tasks without requiring any specific adaptation. In this work, we systematically analyse whether VLMs can be used reliably for image privacy classification in a zero-shot setup. We evaluate and compare the classification performance of three open-source VLMs against purposely built models on two public, standard benchmarks. We assess robustness to image degradations caused by perturbations such as compression, light variations, and random noise, and analyse inference speed and parameter count to deploy VLMs in privacy-aware document processing pipelines. Our results show that large VLMs are robust to input perturbations but are less accurate (and much slower) than smaller privacy models. Scaling alone is not sufficient, highlighting the advantages of models specifically designed for privacy classification. --- ### Smart navigation of a gravity-driven glider with adjustable centre-of-mass Authors: X. Jiang, J. Qiu, K. Gustavsson, B. Mehlig, L. Zhao Published: 2025-10-10 Categories: physics.flu-dyn, cs.LG arXiv: https://arxiv.org/abs/2510.09250v1 Artificial gliders are designed to disperse as they settle through a fluid, requiring precise navigation to reach target locations. We show that a compact glider settling in a viscous fluid can navigate by dynamically adjusting its centre-of-mass. Using fully resolved direct numerical simulations (DNS) and reinforcement learning, we find two optimal navigation strategies that allow the glider to reach its target location accurately. These strategies depend sensitively on how the glider interacts with the surrounding fluid. The nature of this interaction changes as the particle Reynolds number Re$_p$ changes. Our results explain how the optimal strategy depends on Re$_p$. At large Re$_p$, the glider learns to tumble rapidly by moving its centre-of-mass as its orientation changes. This generates a large horizontal inertial lift force, which allows the glider to travel far. At small Re$_p$, by contrast, high viscosity hinders tumbling. In this case, the glider learns to adjust its centre-of-mass so that it settles with a steady, inclined orientation that results in a horizontal viscous force. The horizontal range is much smaller than for large Re$_p$, because this viscous force is much smaller than the inertial lift force at large Re$_p$. *These authors contributed equally. --- ### Application of Deep Reinforcement Learning to At-the-Money S&P 500 Options Hedging Authors: Zofia Bracha, Paweł Sakowski, Jakub Michańków Published: 2025-10-10 Categories: q-fin.CP, cs.LG, q-fin.PR arXiv: https://arxiv.org/abs/2510.09247v1 This paper explores the application of deep Q-learning to hedging at-the-money options on the S\&P~500 index. We develop an agent based on the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm, trained to simulate hedging decisions without making explicit model assumptions on price dynamics. The agent was trained on historical intraday prices of S\&P~500 call options across years 2004--2024, using a single time series of six predictor variables: option price, underlying asset price, moneyness, time to maturity, realized volatility, and current hedge position. A walk-forward procedure was applied for training, which led to nearly 17~years of out-of-sample evaluation. The performance of the deep reinforcement learning (DRL) agent is benchmarked against the Black--Scholes delta-hedging strategy over the same period. We assess both approaches using metrics such as annualized return, volatility, information ratio, and Sharpe ratio. To test the models' adaptability, we performed simulations across varying market conditions and added constraints such as transaction costs and risk-awareness penalties. Our results show that the DRL agent can outperform traditional hedging methods, particularly in volatile or high-cost environments, highlighting its robustness and flexibility in practical trading contexts. While the agent consistently outperforms delta-hedging, its performance deteriorates when the risk-awareness parameter is higher. We also observed that the longer the time interval used for volatility estimation, the more stable the results. --- ### A PCA-based Data Prediction Method Authors: Peteris Daugulis, Vija Vagale, Emiliano Mancini, Filippo Castiglione Published: 2025-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.09246v1 The problem of choosing appropriate values for missing data is often encountered in the data science. We describe a novel method containing both traditional mathematics and machine learning elements for prediction (imputation) of missing data. This method is based on the notion of distance between shifted linear subspaces representing the existing data and candidate sets. The existing data set is represented by the subspace spanned by its first principal components. Solutions for the case of the Euclidean metric are given. --- ### Fundamentals of Building Autonomous LLM Agents Authors: Victor de Lamo Castrillo, Habtom Kahsay Gidey, Alexander Lenz, Alois Knoll Published: 2025-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.09244v1 This paper reviews the architecture and implementation methods of agents powered by large language models (LLMs). Motivated by the limitations of traditional LLMs in real-world tasks, the research aims to explore patterns to develop "agentic" LLMs that can automate complex tasks and bridge the performance gap with human capabilities. Key components include a perception system that converts environmental percepts into meaningful representations; a reasoning system that formulates plans, adapts to feedback, and evaluates actions through different techniques like Chain-of-Thought and Tree-of-Thought; a memory system that retains knowledge through both short-term and long-term mechanisms; and an execution system that translates internal decisions into concrete actions. This paper shows how integrating these systems leads to more capable and generalized software bots that mimic human cognitive processes for autonomous and intelligent behavior. --- ### CrisiText: A dataset of warning messages for LLM training in emergency communication Authors: Giacomo Gonella, Gian Maria Campedelli, Stefano Menini, Marco Guerini Published: 2025-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.09243v2 Effectively identifying threats and mitigating their potential damage during crisis situations, such as natural disasters or violent attacks, is paramount for safeguarding endangered individuals. To tackle these challenges, AI has been used in assisting humans in emergency situations. Still, the use of NLP techniques remains limited and mostly focuses on classification tasks. The significant potential of timely warning message generation using NLG architectures, however, has been largely overlooked. In this paper we present CrisiText, the first large-scale dataset for the generation of warning messages across 13 different types of crisis scenarios. The dataset contains more than 400,000 warning messages (spanning almost 18,000 crisis situations) aimed at assisting civilians during and after such events. To generate the dataset, we started from existing crisis descriptions and created chains of events related to the scenarios. Each event was then paired with a warning message. The generations follow experts' written guidelines to ensure correct terminology and factuality of their suggestions. Additionally, each message is accompanied by three suboptimal warning types to allow for the study of different NLG approaches. To this end, we conducted a series of experiments comparing supervised fine-tuning setups with preference alignment, zero-shot, and few-shot approaches. We further assessed model performance in out-of-distribution scenarios and evaluated the effectiveness of an automatic post-editor. --- ### Investigating the Impact of Rational Dilated Wavelet Transform on Motor Imagery EEG Decoding with Deep Learning Models Authors: Marco Siino, Giuseppe Bonomo, Rosario Sorbello, Ilenia Tinnirello Published: 2025-10-10 Categories: cs.HC, cs.LG arXiv: https://arxiv.org/abs/2510.09242v1 The present study investigates the impact of the Rational Discrete Wavelet Transform (RDWT), used as a plug-in preprocessing step for motor imagery electroencephalographic (EEG) decoding prior to applying deep learning classifiers. A systematic paired evaluation (with/without RDWT) is conducted on four state-of-the-art deep learning architectures: EEGNet, ShallowConvNet, MBEEG\_SENet, and EEGTCNet. This evaluation was carried out across three benchmark datasets: High Gamma, BCI-IV-2a, and BCI-IV-2b. The performance of the RDWT is reported with subject-wise averages using accuracy and Cohen's kappa, complemented by subject-level analyses to identify when RDWT is beneficial. On BCI-IV-2a, RDWT yields clear average gains for EEGTCNet (+4.44 percentage points, pp; kappa +0.059) and MBEEG\_SENet (+2.23 pp; +0.030), with smaller improvements for EEGNet (+2.08 pp; +0.027) and ShallowConvNet (+0.58 pp; +0.008). On BCI-IV-2b, the enhancements observed are modest yet consistent for EEGNet (+0.21 pp; +0.044) and EEGTCNet (+0.28 pp; +0.077). On HGD, average effects are modest to positive, with the most significant gain observed for MBEEG\_SENet (+1.65 pp; +0.022), followed by EEGNet (+0.76 pp; +0.010) and EEGTCNet (+0.54 pp; +0.008). Inspection of the subject material reveals significant enhancements in challenging recordings (e.g., non-stationary sessions), indicating that RDWT can mitigate localized noise and enhance rhythm-specific information. In conclusion, RDWT is shown to be a low-overhead, architecture-aware preprocessing technique that can yield tangible gains in accuracy and agreement for deep model families and challenging subjects. --- ### Incentivizing Time-Aware Fairness in Data Sharing Authors: Jiangwei Chen, Kieu Thao Nguyen Pham, Rachael Hwee Ling Sim, Arun Verma, Zhaoxuan Wu, Chuan-Sheng Foo, Bryan Kian Hsiang Low Published: 2025-10-10 Categories: cs.LG, cs.GT arXiv: https://arxiv.org/abs/2510.09240v2 In collaborative data sharing and machine learning, multiple parties aggregate their data resources to train a machine learning model with better model performance. However, as the parties incur data collection costs, they are only willing to do so when guaranteed incentives, such as fairness and individual rationality. Existing frameworks assume that all parties join the collaboration simultaneously, which does not hold in many real-world scenarios. Due to the long processing time for data cleaning, difficulty in overcoming legal barriers, or unawareness, the parties may join the collaboration at different times. In this work, we propose the following perspective: As a party who joins earlier incurs higher risk and encourages the contribution from other wait-and-see parties, that party should receive a reward of higher value for sharing data earlier. To this end, we propose a fair and time-aware data sharing framework, including novel time-aware incentives. We develop new methods for deciding reward values to satisfy these incentives. We further illustrate how to generate model rewards that realize the reward values and empirically demonstrate the properties of our methods on synthetic and real-world datasets. --- ### Characterizing 5G User Throughput via Uncertainty Modeling and Crowdsourced Measurements Authors: Javier Albert-Smet, Zoraida Frias, Luis Mendo, Sergio Melones, Eduardo Yraola Published: 2025-10-10 Categories: cs.NI, cs.LG arXiv: https://arxiv.org/abs/2510.09239v1 Characterizing application-layer user throughput in next-generation networks is increasingly challenging as the higher capacity of the 5G Radio Access Network (RAN) shifts connectivity bottlenecks towards deeper parts of the network. Traditional methods, such as drive tests and operator equipment counters, are costly, limited, or fail to capture end-to-end (E2E) Quality of Service (QoS) and its variability. In this work, we leverage large-scale crowdsourced measurements-including E2E, radio, contextual and network deployment features collected by the user equipment (UE)-to propose an uncertainty-aware and explainable approach for downlink user throughput estimation. We first validate prior 4G methods, improving R^2 by 8.7%, and then extend them to 5G NSA and 5G SA, providing the first benchmarks for 5G crowdsourced datasets. To address the variability of throughput, we apply NGBoost, a model that outputs both point estimates and calibrated confidence intervals, representing its first use in the field of computer communications. Finally, we use the proposed model to analyze the evolution from 4G to 5G SA, and show that throughput bottlenecks move from the RAN to transport and service layers, as seen by E2E metrics gaining importance over radio-related features. --- ### Diagnosing Shoulder Disorders Using Multimodal Large Language Models and Consumer-Grade Cameras Authors: Jindong Hong, Wencheng Zhang, Shiqin Qiao, Jianhai Chen, Jianing Qiu, Chuanyang Zheng, Qian Xu, Yun Ji, Qianyue Wen, Weiwei Sun, Hao Li, Huizhen Li et al. Published: 2025-10-10 Categories: cs.CV, cs.AI, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2510.09230v1 Shoulder disorders, such as frozen shoulder (a.k.a., adhesive capsulitis), are common conditions affecting the health of people worldwide, and have a high incidence rate among the elderly and workers engaged in repetitive shoulder tasks. In regions with scarce medical resources, achieving early and accurate diagnosis poses significant challenges, and there is an urgent need for low-cost and easily scalable auxiliary diagnostic solutions. This research introduces videos captured by consumer-grade devices as the basis for diagnosis, reducing the cost for users. We focus on the innovative application of Multimodal Large Language Models (MLLMs) in the preliminary diagnosis of shoulder disorders and propose a Hybrid Motion Video Diagnosis framework (HMVDx). This framework divides the two tasks of action understanding and disease diagnosis, which are respectively completed by two MLLMs. In addition to traditional evaluation indicators, this work proposes a novel metric called Usability Index by the logical process of medical decision-making (action recognition, movement diagnosis, and final diagnosis). This index evaluates the effectiveness of MLLMs in the medical field from the perspective of the entire medical diagnostic pathway, revealing the potential value of low-cost MLLMs in medical applications for medical practitioners. In experimental comparisons, the accuracy of HMVDx in diagnosing shoulder joint injuries has increased by 79.6\% compared with direct video diagnosis, a significant technical contribution to future research on the application of MLLMs for video understanding in the medical field. --- ### Clear Roads, Clear Vision: Advancements in Multi-Weather Restoration for Smart Transportation Authors: Vijay M. Galshetwar, Praful Hambarde, Prashant W. Patil, Akshay Dudhane, Sachin Chaudhary Published: 2025-10-10 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2510.09228v2 Adverse weather conditions such as haze, rain, and snow significantly degrade the quality of images and videos, posing serious challenges to intelligent transportation systems (ITS) that rely on visual input. These degradations affect critical applications including autonomous driving, traffic monitoring, and surveillance. This survey presents a comprehensive review of image and video restoration techniques developed to mitigate weather-induced visual impairments. We categorize existing approaches into traditional prior-based methods and modern data-driven models, including CNNs, transformers, diffusion models, and emerging vision-language models (VLMs). Restoration strategies are further classified based on their scope: single-task models, multi-task/multi-weather systems, and all-in-one frameworks capable of handling diverse degradations. In addition, we discuss day and night time restoration challenges, benchmark datasets, and evaluation protocols. The survey concludes with an in-depth discussion on limitations in current research and outlines future directions such as mixed/compound-degradation restoration, real-time deployment, and agentic AI frameworks. This work aims to serve as a valuable reference for advancing weather-resilient vision systems in smart transportation environments. Lastly, to stay current with rapid advancements in this field, we will maintain regular updates of the latest relevant papers and their open-source implementations at https://github.com/ChaudharyUPES/A-comprehensive-review-on-Multi-weather-restoration --- ### RegexPSPACE: A Benchmark for Evaluating LLM Reasoning on PSPACE-complete Regex Problems Authors: Hyundong Jin, Joonghyuk Hahn, Yo-Sub Han Published: 2025-10-10 Categories: cs.AI, cs.FL arXiv: https://arxiv.org/abs/2510.09227v1 Large language models (LLMs) show strong performance across natural language processing (NLP), mathematical reasoning, and programming, and recent large reasoning models (LRMs) further emphasize explicit reasoning. Yet their computational limits, particularly spatial complexity constrained by finite context windows, remain poorly understood. While recent works often focus on problems within the NP complexity class, we push the boundary by introducing a novel benchmark grounded in two PSPACE-complete regular expression (regex) problems: equivalence decision (RegexEQ) and minimization (RegexMin). PSPACE-complete problems serve as a more rigorous standard for assessing computational capacity, as their solutions require massive search space exploration. We perform a double-exponential space exploration to construct a labeled dataset of over a million regex instances with a sound filtering process to build the benchmark. We conduct extensive evaluations on 6 LLMs and 5 LRMs of varying scales, revealing common failure patterns such as verbosity and repetition. With its well-defined structure and quantitative evaluation metrics, this work presents the first empirical investigation into the spatial computational limitations of LLMs and LRMs, offering a new framework for evaluating their advanced reasoning capabilities. Our code is available at https://github.com/hyundong98/RegexPSPACE . --- ### Prime Implicant Explanations for Reaction Feasibility Prediction Authors: Klaus Weinbauer, Tieu-Long Phan, Peter F. Stadler, Thomas Gärtner, Sagar Malhotra Published: 2025-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.09226v1 Machine learning models that predict the feasibility of chemical reactions have become central to automated synthesis planning. Despite their predictive success, these models often lack transparency and interpretability. We introduce a novel formulation of prime implicant explanations--also known as minimally sufficient reasons--tailored to this domain, and propose an algorithm for computing such explanations in small-scale reaction prediction tasks. Preliminary experiments demonstrate that our notion of prime implicant explanations conservatively captures the ground truth explanations. That is, such explanations often contain redundant bonds and atoms but consistently capture the molecular attributes that are essential for predicting reaction feasibility. --- ### Unsupervised lexicon learning from speech is limited by representations rather than clustering Authors: Danel Slabbert, Simon Malan, Herman Kamper Published: 2025-10-10 Categories: eess.AS, cs.CL, cs.SD arXiv: https://arxiv.org/abs/2510.09225v2 Zero-resource word segmentation and clustering systems aim to tokenise speech into word-like units without access to text labels. Despite progress, the induced lexicons are still far from perfect. In an idealised setting with gold word boundaries, we ask whether performance is limited by the representation of word segments, or by the clustering methods that group them into word-like types. We combine a range of self-supervised speech features (continuous/discrete, frame/word-level) with different clustering methods (K-means, hierarchical, graph-based) on English and Mandarin data. The best system uses graph clustering with dynamic time warping on continuous features. Faster alternatives use graph clustering with cosine distance on averaged continuous features or edit distance on discrete unit sequences. Through controlled experiments that isolate either the representations or the clustering method, we demonstrate that representation variability across segments of the same word type -- rather than clustering -- is the primary factor limiting performance. --- ### Tag-Enriched Multi-Attention with Large Language Models for Cross-Domain Sequential Recommendation Authors: Wangyu Wu, Xuhang Chen, Zhenhong Chen, Jing-En Jiang, Kim-Fung Tsang, Xiaowei Huang, Fei Ma, Jimin Xiao Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09224v2 Cross-Domain Sequential Recommendation (CDSR) plays a crucial role in modern consumer electronics and e-commerce platforms, where users interact with diverse services such as books, movies, and online retail products. These systems must accurately capture both domain-specific and cross-domain behavioral patterns to provide personalized and seamless consumer experiences. To address this challenge, we propose \textbf{TEMA-LLM} (\textit{Tag-Enriched Multi-Attention with Large Language Models}), a practical and effective framework that integrates \textit{Large Language Models (LLMs)} for semantic tag generation and enrichment. Specifically, TEMA-LLM employs LLMs to assign domain-aware prompts and generate descriptive tags from item titles and descriptions. The resulting tag embeddings are fused with item identifiers as well as textual and visual features to construct enhanced item representations. A \textit{Tag-Enriched Multi-Attention} mechanism is then introduced to jointly model user preferences within and across domains, enabling the system to capture complex and evolving consumer interests. Extensive experiments on four large-scale e-commerce datasets demonstrate that TEMA-LLM consistently outperforms state-of-the-art baselines, underscoring the benefits of LLM-based semantic tagging and multi-attention integration for consumer-facing recommendation systems. The proposed approach highlights the potential of LLMs to advance intelligent, user-centric services in the field of consumer electronics. --- ### Comparing Knowledge Source Integration Methods for Optimizing Healthcare Knowledge Fusion in Rescue Operation Authors: Mubaris Nadeem, Madjid Fathi Published: 2025-10-10 Categories: cs.AI arXiv: https://arxiv.org/abs/2510.09223v1 In the field of medicine and healthcare, the utilization of medical expertise, based on medical knowledge combined with patients' health information is a life-critical challenge for patients and health professionals. The within-laying complexity and variety form the need for a united approach to gather, analyze, and utilize existing knowledge of medical treatments, and medical operations to provide the ability to present knowledge for the means of accurate patient-driven decision-making. One way to achieve this is the fusion of multiple knowledge sources in healthcare. It provides health professionals the opportunity to select from multiple contextual aligned knowledge sources which enables the support for critical decisions. This paper presents multiple conceptual models for knowledge fusion in the field of medicine, based on a knowledge graph structure. It will evaluate, how knowledge fusion can be enabled and presents how to integrate various knowledge sources into the knowledge graph for rescue operations. --- ### FM-IRL: Flow-Matching for Reward Modeling and Policy Regularization in Reinforcement Learning Authors: Zhenglin Wan, Jingxuan Wu, Xingrui Yu, Chubin Zhang, Mingcong Lei, Bo An, Ivor Tsang Published: 2025-10-10 Categories: cs.LG arXiv: https://arxiv.org/abs/2510.09222v3 Flow Matching (FM) has shown remarkable ability in modeling complex distributions and achieves strong performance in offline imitation learning for cloning expert behaviors. However, despite its behavioral cloning expressiveness, FM-based policies are inherently limited by their lack of environmental interaction and exploration. This leads to poor generalization in unseen scenarios beyond the expert demonstrations, underscoring the necessity of online interaction with environment. Unfortunately, optimizing FM policies via online interaction is challenging and inefficient due to instability in gradient computation and high inference costs. To address these issues, we propose to let a student policy with simple MLP structure explore the environment and be online updated via RL algorithm with a reward model. This reward model is associated with a teacher FM model, containing rich information of expert data distribution. Furthermore, the same teacher FM model is utilized to regularize the student policy's behavior to stabilize policy learning. Due to the student's simple architecture, we avoid the gradient instability of FM policies and enable efficient online exploration, while still leveraging the expressiveness of the teacher FM model. Extensive experiments show that our approach significantly enhances learning efficiency, generalization, and robustness, especially when learning from suboptimal expert data. --- ### IRIS: An Iterative and Integrated Framework for Verifiable Causal Discovery in the Absence of Tabular Data Authors: Tao Feng, Lizhen Qu, Niket Tandon, Gholamreza Haffari Published: 2025-10-10 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2510.09217v1 Causal discovery is fundamental to scientific research, yet traditional statistical algorithms face significant challenges, including expensive data collection, redundant computation for known relations, and unrealistic assumptions. While recent LLM-based methods excel at identifying commonly known causal relations, they fail to uncover novel relations. We introduce IRIS (Iterative Retrieval and Integrated System for Real-Time Causal Discovery), a novel framework that addresses these limitations. Starting with a set of initial variables, IRIS automatically collects relevant documents, extracts variables, and uncovers causal relations. Our hybrid causal discovery method combines statistical algorithms and LLM-based methods to discover known and novel causal relations. In addition to causal discovery on initial variables, the missing variable proposal component of IRIS identifies and incorporates missing variables to expand the causal graphs. Our approach enables real-time causal discovery from only a set of initial variables without requiring pre-existing datasets. --- ### Stable Video Infinity: Infinite-Length Video Generation with Error Recycling Authors: Wuyang Li, Wentao Pan, Po-Chien Luan, Yang Gao, Alexandre Alahi Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09212v1 We propose Stable Video Infinity (SVI) that is able to generate infinite-length videos with high temporal consistency, plausible scene transitions, and controllable streaming storylines. While existing long-video methods attempt to mitigate accumulated errors via handcrafted anti-drifting (e.g., modified noise scheduler, frame anchoring), they remain limited to single-prompt extrapolation, producing homogeneous scenes with repetitive motions. We identify that the fundamental challenge extends beyond error accumulation to a critical discrepancy between the training assumption (seeing clean data) and the test-time autoregressive reality (conditioning on self-generated, error-prone outputs). To bridge this hypothesis gap, SVI incorporates Error-Recycling Fine-Tuning, a new type of efficient training that recycles the Diffusion Transformer (DiT)'s self-generated errors into supervisory prompts, thereby encouraging DiT to actively identify and correct its own errors. This is achieved by injecting, collecting, and banking errors through closed-loop recycling, autoregressively learning from error-injected feedback. Specifically, we (i) inject historical errors made by DiT to intervene on clean inputs, simulating error-accumulated trajectories in flow matching; (ii) efficiently approximate predictions with one-step bidirectional integration and calculate errors with residuals; (iii) dynamically bank errors into replay memory across discretized timesteps, which are resampled for new input. SVI is able to scale videos from seconds to infinite durations with no additional inference cost, while remaining compatible with diverse conditions (e.g., audio, skeleton, and text streams). We evaluate SVI on three benchmarks, including consistent, creative, and conditional settings, thoroughly verifying its versatility and state-of-the-art role. --- ### DICE: Structured Reasoning in LLMs through SLM-Guided Chain-of-Thought Correction Authors: Yiqi Li, Yusheng Liao, Zhe Chen, Yanfeng Wang, Yu Wang Published: 2025-10-10 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2510.09211v2 When performing reasoning tasks with user-specific requirements, such as strict output formats, large language models (LLMs) often prioritize reasoning over adherence to detailed instructions. Fine-tuning LLMs on supervised datasets to address this is impractical due to high computational costs and limited parameter access. To tackle this, we propose DICE, a lightweight framework that guides small language models (SLMs) to refine LLMs' outputs through chain-of-thought (CoT) correction. DICE decouples the process by first prompting LLMs to generate natural language responses, then using trained SLMs to analyze and refine these outputs to meet structured output specifications. This framework preserves LLMs' broad knowledge and reasoning capabilities while ensuring the outputs conform to user demands. Specifically, DICE first constructs structured CoT adaptation datasets via a two-stage method and subsequently applies a dual-tuning strategy to fine-tune SLMs for generating structured outputs in an analyze-then-answer pattern. Experiments demonstrate that DICE improves the average format accuracy and content correctness of LLM outputs by 35.4\% and 29.4\%, respectively, achieving state-of-the-art (SOTA) performance over other competitive baselines. --- ### Provable Watermarking for Data Poisoning Attacks Authors: Yifan Zhu, Lijia Yu, Xiao-Shan Gao Published: 2025-10-10 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2510.09210v1 In recent years, data poisoning attacks have been increasingly designed to appear harmless and even beneficial, often with the intention of verifying dataset ownership or safeguarding private data from unauthorized use. However, these developments have the potential to cause misunderstandings and conflicts, as data poisoning has traditionally been regarded as a security threat to machine learning systems. To address this issue, it is imperative for harmless poisoning generators to claim ownership of their generated datasets, enabling users to identify potential poisoning to prevent misuse. In this paper, we propose the deployment of watermarking schemes as a solution to this challenge. We introduce two provable and practical watermarking approaches for data poisoning: {\em post-poisoning watermarking} and {\em poisoning-concurrent watermarking}. Our analyses demonstrate that when the watermarking length is $Θ(\sqrt{d}/ε_w)$ for post-poisoning watermarking, and falls within the range of $Θ(1/ε_w^2)$ to $O(\sqrt{d}/ε_p)$ for poisoning-concurrent watermarking, the watermarked poisoning dataset provably ensures both watermarking detectability and poisoning utility, certifying the practicality of watermarking under data poisoning attacks. We validate our theoretical findings through experiments on several attacks, models, and datasets. --- ### 3D Reconstruction from Transient Measurements with Time-Resolved Transformer Authors: Yue Li, Shida Sun, Yu Hong, Feihu Xu, Zhiwei Xiong Published: 2025-10-10 Categories: cs.CV, eess.IV arXiv: https://arxiv.org/abs/2510.09205v1 Transient measurements, captured by the timeresolved systems, are widely employed in photon-efficient reconstruction tasks, including line-of-sight (LOS) and non-line-of-sight (NLOS) imaging. However, challenges persist in their 3D reconstruction due to the low quantum efficiency of sensors and the high noise levels, particularly for long-range or complex scenes. To boost the 3D reconstruction performance in photon-efficient imaging, we propose a generic Time-Resolved Transformer (TRT) architecture. Different from existing transformers designed for high-dimensional data, TRT has two elaborate attention designs tailored for the spatio-temporal transient measurements. Specifically, the spatio-temporal self-attention encoders explore both local and global correlations within transient data by splitting or downsampling input features into different scales. Then, the spatio-temporal cross attention decoders integrate the local and global features in the token space, resulting in deep features with high representation capabilities. Building on TRT, we develop two task-specific embodiments: TRT-LOS for LOS imaging and TRT-NLOS for NLOS imaging. Extensive experiments demonstrate that both embodiments significantly outperform existing methods on synthetic data and real-world data captured by different imaging systems. In addition, we contribute a large-scale, high-resolution synthetic LOS dataset with various noise levels and capture a set of real-world NLOS measurements using a custom-built imaging system, enhancing the data diversity in this field. Code and datasets are available at https://github.com/Depth2World/TRT. --- ### Flow-Opt: Scalable Centralized Multi-Robot Trajectory Optimization with Flow Matching and Differentiable Optimization Authors: Simon Idoko, Prajyot Jadhav, Arun Kumar Singh Published: 2025-10-10 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2510.09204v4 Centralized trajectory optimization in the joint space of multiple robots allows access to a larger feasible space that can result in smoother trajectories, especially while planning in tight spaces. Unfortunately, it is often computationally intractable beyond a very small swarm size. In this paper, we propose Flow-Opt, a learning-based approach towards improving the computational tractability of centralized multi-robot trajectory optimization. Specifically, we reduce the problem to first learning a generative model to sample different candidate trajectories and then using a learned Safety-Filter(SF) to ensure fast inference-time constraint satisfaction. We propose a flow-matching model with a diffusion transformer (DiT) augmented with permutation invariant robot position and map encoders as the generative model. We develop a custom solver for our SF and equip it with a neural network that predicts context-specific initialization. The initialization network is trained in a self-supervised manner, taking advantage of the differentiability of the SF solver. We advance the state-of-the-art in the following respects. First, we show that we can generate trajectories of tens of robots in cluttered environments in a few tens of milliseconds. This is several times faster than existing centralized optimization approaches. Moreover, our approach also generates smoother trajectories orders of magnitude faster than competing baselines based on diffusion models. Second, each component of our approach can be batched, allowing us to solve a few tens of problem instances in a fraction of a second. We believe this is a first such result; no existing approach provides such capabilities. Finally, our approach can generate a diverse set of trajectories between a given set of start and goal locations, which can capture different collision-avoidance behaviors. --- ### Cattle-CLIP: A Multimodal Framework for Cattle Behaviour Recognition from Video Authors: Huimin Liu, Jing Gao, Daria Baran, AxelX Montout, Neill W Campbell, Andrew W Dowsey Published: 2025-10-10 Categories: cs.CV arXiv: https://arxiv.org/abs/2510.09203v2 Robust behaviour recognition in real-world farm environments remains challenging due to several data-related limitations, including the scarcity of well-annotated livestock video datasets and the substantial domain gap between large-scale pre-training corpora and agricultural surveillance footage. To address these challenges, we propose Cattle-CLIP, a domain-adaptive vision-language framework that reformulates cattle behaviour recognition as cross-modal semantic alignment rather than purely visual classification. Instead of directly fine-tuning visual backbones, Cattle-CLIP incorporates a temporal integration module to extend image-level contrastive pre-training to video-based behaviour understanding, enabling consistent semantic alignment across time. To mitigate the distribution shift between web-scale image-text data used for the pre-trained model and real-world cattle surveillance footage, we further introduce tailored augmentation strategies and specialised behaviour prompts. Furthermore, we construct CattleBehaviours6, a curated and behaviour-consistent video dataset comprising 1905 annotated clips across six indoor behaviours to support model training and evaluation. Beyond serving as a benchmark for our proposed method, the dataset provides a standardised ethogram definition, offering a practical resource for future research in livestock behaviour analysis. Cattle-CLIP is evaluated under both fully-supervised and few-shot learning scenarios, with a particular focus on data-scarce behaviour recognition, an important yet under-explored goal in livestock monitoring. Experiments show that Cattle-CLIP achieves 96.1% overall accuracy across six behaviours in supervised settings, with near-perfect recall for feeding, drinking and standing-ruminating behaviours, and demonstrates robust generalisation with limited data in few-shot scenarios. --- ### Multimodal Prompt Optimization: Why Not Leverage Multiple Modalities for MLLMs Authors: Yumin Choi, Dongki Kim, Jinheon Baek, Sung Ju Hwang Published: 2025-10-10 Categories: cs.LG, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2510.09201v2 Large Language Models (LLMs) have shown remarkable success, and their multimodal expansions (MLLMs) further unlock capabilities spanning images, videos, and other modalities beyond text. However, despite this shift, prompt optimization approaches, designed to reduce the burden of manual prompt crafting while maximizing performance, remain confined to text, ultimately limiting the full potential of MLLMs. Motivated by this gap, we introduce the new problem of multimodal prompt optimization, which expands the prior definition of prompt optimization to the multimodal space defined by the pairs of textual and non-textual prompts. To tackle this problem, we then propose the Multimodal Prompt Optimizer (MPO), a unified framework that not only performs the joint optimization of multimodal prompts through alignment-preserving updates but also guides the selection process of candidate prompts by leveraging earlier evaluations as priors in a Bayesian-based selection strategy. Through extensive experiments across diverse modalities that go beyond text, such as images, videos, and even molecules, we demonstrate that MPO outperforms leading text-only optimization methods, establishing multimodal prompt optimization as a crucial step to realizing the potential of MLLMs. --- ### Towards Safer and Understandable Driver Intention Prediction Authors: Mukilan Karuppasamy, Shankar Gangisetty, Shyam Nandan Rai, Carlo Masone, C V Jawahar Published: 2025-10-10 Categories: cs.CV, cs.AI, cs.HC arXiv: https://arxiv.org/abs/2510.09200v1 Autonomous driving (AD) systems are becoming increasingly capable of handling complex tasks, mainly due to recent advances in deep learning and AI. As interactions between autonomous systems and humans increase, the interpretability of decision-making processes in driving systems becomes increasingly crucial for ensuring safe driving operations. Successful human-machine interaction requires understanding the underlying representations of the environment and the driving task, which remains a significant challenge in deep learning-based systems. To address this, we introduce the task of interpretability in maneuver prediction before they occur for driver safety, i.e., driver intent prediction (DIP), which plays a critical role in AD systems. To foster research in interpretable DIP, we curate the eXplainable Driving Action Anticipation Dataset (DAAD-X), a new multimodal, ego-centric video dataset to provide hierarchical, high-level textual explanations as causal reasoning for the driver's decisions. These explanations are derived from both the driver's eye-gaze and the ego-vehicle's perspective. Next, we propose Video Concept Bottleneck Model (VCBM), a framework that generates spatio-temporally coherent explanations inherently, without relying on post-hoc techniques. Finally, through extensive evaluations of the proposed VCBM on the DAAD-X dataset, we demonstrate that transformer-based models exhibit greater interpretability than conventional CNN-based models. Additionally, we introduce a multilabel t-SNE visualization technique to illustrate the disentanglement and causal correlation among multiple explanations. Our data, code and models are available at: https://mukil07.github.io/VCBM.github.io/ --- ### Augmented data and neural networks for robust epidemic forecasting: application to COVID-19 in Italy Authors: Giacomo Dimarco, Federica Ferrarese, Lorenzo Pareschi Published: 2025-10-10 Categories: math.NA, cs.LG arXiv: https://arxiv.org/abs/2510.09192v2 In this work, we propose a data augmentation strategy aimed at improving the training phase of neural networks and, consequently, the accuracy of their predictions. Our approach relies on generating synthetic data through a suitable compartmental model combined with the incorporation of uncertainty. The available data are then used to calibrate the model, which is further integrated with deep learning techniques to produce additional synthetic data for training. The results show that neural networks trained on these augmented datasets exhibit significantly improved predictive performance. We focus in particular on two different neural network architectures: Physics-Informed Neural Networks (PINNs) and Nonlinear Autoregressive (NAR) models. The NAR approach proves especially effective for short-term forecasting, providing accurate quantitative estimates by directly learning the dynamics from data and avoiding the additional computational cost of embedding physical constraints into the training. In contrast, PINNs yield less accurate quantitative predictions but capture the qualitative long-term behavior of the system, making them more suitable for exploring broader dynamical trends. Numerical simulations of the second phase of the COVID-19 pandemic in the Lombardy region (Italy) validate the effectiveness of the proposed approach. ---