# Research Papers — 2026-06 · bundle page 45 of 260 > Full-text markdown of papers #2201–2250 of 12993 published in 2026-06, newest first. AI-readable mirrors of the full 12993-paper 2026-06 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-06/page/45 [← Previous bundle](https://aipod.dev/papers/archive/2026-06/page/44.md) [Next bundle →](https://aipod.dev/papers/archive/2026-06/page/46.md) ### RolloutPipe: Overlapping Pipelined Rollout and Training in Disaggregated On-Policy LLM Reinforcement Learning Authors: Rongjian Chen, Jianmin Hu, Kejiang Ye, Minxian Xu Published: 2026-06-25 Categories: cs.DC, cs.LG arXiv: https://arxiv.org/abs/2606.26997v2 Large language model (LLM) post-training for reasoning increasingly relies on reinforcement learning with verifiable rewards (RLVR), where models learn from ground-truth feedback on mathematical, logical, and scientific tasks. To enable flexible resource allocation and support heterogeneous training setups, modern RLVR systems adopt disaggregated architectures that decouple rollout generation and policy training across independent GPU pools. However, existing synchronous on-policy GRPO (Group Relative Policy Optimization) RLVR systems finish an entire rollout before starting training, leaving the trainer GPU pool idle while rollout is still ongoing. Asynchronous RL pipelines overlap the two stages, but at the cost of training on stale data. To address these challenges, we propose RolloutPipe, a post-training framework for disaggregated RLVR systems, which turns the fixed-weight rollout into a complete-group pipeline where trainable groups move to the trainer while later groups are still being generated. RolloutPipe achieves this through two techniques including complete-group pipelining (CGP) and frontier-group dispatch (FGD). CGP dispatches each trainable complete group to the trainer FIFO as soon as group materialization finishes, and FGD is an admission policy on the Rollout node that first admits requests for the frontier groups needed to form the next training batch, so that trainer-ready groups arrive earlier and more steadily. The design starts training before the rollout completes while maintaining on-policy correctness. Evaluated on Qwen3-1.7B across four reasoning and science benchmarks and twelve rollout settings, RolloutPipe shortens the rollout-to-train-end time by 30.7%-42.3%, and lowers the trainer waiting ratio by 37%-76% compared to Slime, a state-of-the-art rollout and training system. --- ### Event-Aware Instructed Assistant for Referring Video Segmentation Authors: Jinyu Liu, Henghui Ding, Shuting He, Yu-Gang Jiang Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26994v1 Existing referring video segmentation methods often treat a video as a single event consisting of multiple images, overlooking the fact that a video typically contains multiple distinct events. Under such a mechanism, the model needs to directly understand all the complex content in the video and text, which can easily lead to confusion and hallucinations. To address this issue, we propose to decompose a video to a set of simple events by learnable Event Query, and understand complex video content in an event-by-event, easy-to-understand manner. This is based on the observation that natural language expressions often divide a video into distinct, text-related segments, each representing a separate event within a compound event. We introduce EVIS, an Event-Aware Video Instructed Segmentation Assistant, which utilizes text-guided Event Queries to partition a video into simple events, extracting event-aware visual-text features to achieve a hierarchical understanding of the video. Additionally, we propose Object-Pixel-Hybrid Learning, which enables the MLLMs to track targets in long-term videos by integrating fine-grained pixel features with prior object queries. Extensive experimental results on 5 public benchmarks demonstrate EVIS's strong performance in addressing the referring video segmentation task. --- ### Enabling self-supervised learned primal dual with Noise2Inverse Authors: Antti Sällinen, Siiri Rautio, Santeri Kaupinmäki, Andreas Hauptmann Published: 2026-06-25 Categories: eess.IV, cs.LG, math.OC arXiv: https://arxiv.org/abs/2606.26991v1 X-ray computed tomography reconstruction is an ill-posed inverse problem, particularly in low-dose and sparse-angle settings where measurements are noisy and incomplete. While learned reconstruction methods such as the Learned Primal-Dual algorithm achieve strong performance, they typically rely on supervised training with access to ground-truth data, which is often unavailable in practice. In this work, we propose a self-supervised reconstruction method by extending the Noise2Inverse framework to the Learned Primal-Dual algorithm. The resulting approach, called Noise2Inverse Learned Primal-Dual (N2I-LPD), enables training of a learned iterative reconstruction operator without ground-truth images by exploiting the statistical independence of noise in distinct measurements with respect to angular rotation of the CT-scan. We compare the proposed method with classical reconstruction methods, as well as neural network-based approaches such as a U-Net trained within the same N2I framework. The results demonstrate that N2I-LPD achieves improved reconstruction quality, highlighting the potential of combining learned reconstruction operators with self-supervised training strategies for practical CT imaging scenarios where ground-truth data is unavailable. --- ### Decision-Aligned Evaluation of Uncertainty Quantification Authors: Annika Schneider, Tommy Rochussen, Joshua Stiller, Vincent Fortuin Published: 2026-06-25 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2606.26990v1 Uncertainty estimates in machine learning are typically evaluated using generic metrics such as the negative log-likelihood and expected calibration error, yet good performance on such metrics does not necessarily imply high utility in downstream decisions. We introduce decision-alignment, a criterion that reveals which evaluation metrics meaningfully align with downstream utilities. Applying this framework, we show that many widely used uncertainty metrics are either misaligned with common decision problems or encode pathological prior beliefs about the downstream task. We then propose prior-weighted utility metrics, a special class of proper scoring rules that provides decision-aligned uncertainty evaluation. Across benchmark experiments and real-world case studies, our metrics consistently align with realized decision utility, while conventional metrics do not. Our results surface flaws in the current UQ evaluation protocol and offer a principled extension of existing metrics toward decision-relevant UQ evaluation. --- ### Where Do Models Find Happiness? Emotion Vectors in Open-Source LLMs Authors: Sinie van der Ben, Raphaël Baur, Yannick Metz, Mennatallah El-Assady Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.26987v1 Recent work identified emotion vectors in Claude Sonnet 4.5, which are internal representations that encode emotion concepts, causally influence behavior, and exhibit geometry mirroring human psychological structure. We test the generality of these findings in two open-weight models, Apertus-8B-Instruct-2509 and Gemma-4-E4B-it, extracting emotion contrast vectors across all layers, using two model-generated corpora. We recover valence geometry for both models, with peak PC1--valence correlations of $r = 0.76$ and $r = 0.83$, approaching the $r = 0.81$ reported for Claude.Beyond replication, we observe notable differences in how valence representations emerge across model depth. In Gemma-4-E4B-it, valence is strongly encoded in early layers but collapses towards later layers, whereas Apertus-8B-Instruct-2509 exhibits the opposite pattern, with valence representations absent in early layers, but emerging at mid depths. Arousal encoding, in contrast, is sensitive to the extraction corpus: both models show stronger PC2--arousal alignment with Gemma-generated stories ($r$ up to $0.45$) than Apertus-generated ones ($r \leq 0.21$), suggesting arousal-relevant cues are unevenly distributed across generated corpora. We open-source our experiment code and dataset for reproducible investigation of emotion representations across language model architectures. --- ### ReaORE: Reasoning-Guided Progressive Open Relation Extraction Empowered by Large Reasoning Models Authors: Xin Lin, Liang Zhang, Guoqi Ma, Hongyao Tu, Jinsong Su Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.26986v1 Open Relation Extraction (OpenRE) requires a model to extract unseen relations between head and tail entities from unstructured text for real-world applications. The core challenge of OpenRE lies in achieving reliable generalization to unseen relation types. Current OpenRE approaches either employ clustering techniques, which cannot generate relation labels and suffer from poor generalization, or rely on direct relation label generation via Large Language Models (LLMs), which lack sufficient discriminative capacity to distinguish easily confused relations. To address these limitations, we propose Reasoning-guided progressive OpenRE (ReaORE), a framework for performing relation extraction through coarse-to-fine relation reasoning. Specifically, ReaORE consists of two key stages: (i) relation filtering, which reasons over multiple aspects to understand relations and instances, yielding an initial relation set, and further supplements and filters relations via embedding-based similarity to ensure the target relation is included; (ii) relation prediction, which aims to predict the target relations from the above set via fine-grained comparative reasoning to better distinguish easily confused relations. Extensive experiments on two widely used OpenRE datasets demonstrate that ReaORE outperforms existing baselines. --- ### Unison: Benchmarking Unified Multimodal Models via Synergistic Understanding and Generation Authors: Jinyu Liu, Xincheng Shuai, Henghui Ding, Yu-Gang Jiang Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26984v1 Unified multimodal models capable of both understanding and generation have achieved remarkable strides. However, despite their unified designs, existing evaluations typically assess understanding and generation capabilities in isolation, overlooking the synergy between comprehension and generation. To bridge this gap, we introduce Unison, a comprehensive benchmark comprising 2,169 high-quality unified task samples, designed to evaluate joint understanding and generation in unified multimodal models. Unison offers three key strengths: 1) Comprehensive Dimensions: Unison encompasses internal consistency, understanding-guided generation, generation-guided understanding, and mutual enhancement to enable holistic evaluation. 2) Diagnostic Evaluation: it provides both unified and decoupled tracks for understanding and generation, allowing fine-grained attribution of failure modes and quantitative analysis of the gains from unified modeling. 3) Human Alignment: we also introduce Unison-Judge, an evaluation model well aligned with human judgments to ensure reliable assessment. Based on systematic evaluations of state-of-the-art models on Unison, we uncover critical limitations in current unified multimodal systems and highlight promising directions for future research. Codes, Unison and Unison-Judge are publicly available at https://github.com/FudanCVL/Unison. --- ### Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactions Authors: Abla Bedoui, Ashley L. Greene, Mohammed Cherkaoui Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.26982v1 Large language models (LLMs) are increasingly being integrated into mental health support tools and other psychologically sensitive conversational applications. In such settings, behavioral stability and consistency are important for trustworthy human-AI interaction. However, semantically similar concerns can be presented through different contextual framings, potentially eliciting different model responses. Such framing-sensitive variability may challenge user expectations regarding system behavior and complicate the assessment of AI reliability. While prior studies have primarily examined such effects at the behavioral level, less is known about how framing-related variation is reflected in the internal representations of aligned language models. In this work, we investigate these effects using controlled matched prompts spanning multiple contextual framing conditions across several instruction-tuned model families. Across architectures, framing systematically alters interpretive response tendencies. Layer-wise probing analyses show that behavior-associated information remains decodable throughout transformer depth, with architecture-dependent variation in decoding strength. Moreover, held-out framing probes remained consistently above chance across architectures despite strong lexical baselines. Activation steering experiments further suggest that framing-associated representational directions can partially modulate downstream behavioral outcomes. Finally, these findings indicate that robustness to contextual variation may represent an important consideration when evaluating the consistency and trustworthiness of conversational AI systems deployed in mental-health-oriented interactions. --- ### In-Context Model Predictive Generation: Open-Vocabulary Motion Synthesis from Language Models to Physics Authors: Xiaomeng Fu, Junfan Lin, Yang Liu, Yaowei Wang, Guanbin Li, Liang Lin, Ziliang Chen Published: 2026-06-25 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.26981v1 Synthesizing human motion from textual descriptions is essential for immersive digital applications, yet existing methods face a persistent trade-off between semantic fidelity and physical realism. Large language model (LLM)-based approaches can interpret diverse open-vocabulary instructions and compose high-level action plans, but they often generate motions that violate physical constraints. Physics-aware models improve realism through simulation or control, but they struggle with semantic complexity, fine-grained instructions, and novel concepts. To address this gap, we propose In-Context Model Predictive Generation (ICMPG), a framework that integrates language-model planning with inference-time physical feedback. ICMPG reformulates motion synthesis as a Model Predictive Control (MPC)-like process with two modules. The Context-Aware Motion Generation (CAMG) module uses an LLM as a planner to decompose textual commands and generate candidate motion sequences from motion tokens. The Model Predictive Generation (MPG) module evaluates these candidates through physical simulation and semantic alignment, estimates a composite reward, and selects the best sequence to guide subsequent generation steps. Unlike open-loop generation, this closed-loop refinement enables ICMPG to adapt motions to both the input semantics and the simulated physical environment without task-specific policy retraining. Extensive experiments across standard and zero-shot open-vocabulary settings show that ICMPG generalizes robustly to diverse commands and produces motions that are more physically plausible and semantically faithful than representative baselines on the evaluated benchmarks. The framework bridges semantic interpretation and physical simulation while remaining flexible enough to incorporate different LLM backbones, enabling more versatile and controllable text-driven motion synthesis. --- ### XMSE-Aware Adaptive Empirical Bayes Estimation Authors: Minghao Chen, Jiale Zheng Published: 2026-06-25 Categories: stat.ML, cs.AI, cs.LG, eess.SY, stat.ME arXiv: https://arxiv.org/abs/2606.26975v1 Empirical Bayes (EB) estimators can match the first-order asymptotic risk of maximum likelihood (ML) while behaving very differently at second order: recent excess mean squared error (XMSE) analysis shows that kernel-based EB estimation may be worse than ML when the kernel is poorly aligned with the true parameter. This paper turns that diagnostic into a design principle. We propose an XMSE-aware mixed estimator that interpolates between ML and EB shrinkage. Its fixed-weight XMSE is a scalar quadratic, yielding a closed-form oracle mixing weight that is no worse than both ML and the base EB estimator at the XMSE scale. A plug-in implementation based on finite-sample XMSE approximations is proved consistent, with a second-order oracle regret rate for an interior oracle weight. We further establish a transfer of the regret bound to the fixed-weight risk curve evaluated at the selected weight, a thresholded boundary rule, and extensions to compact kernel families and to finite and growing kernel dictionaries with high-probability oracle bounds. Finite impulse response simulations with SURE-tuned, hard-selection, and trace-corrected baselines, together with the public Silverbox and Cascaded Tanks benchmarks, show that the proposed estimator retains most of the benefit of regularization when it is helpful and retreats toward ML under kernel misspecification, with an identified finite-de analyzed on the benchmarks. --- ### Geometric Gradient Rectification for Safe Open-Set Semi-Supervised Learning Authors: Jiahe Chen, Qian Shao, Qiyuan Chen, Jiaying He, Jintai Chen, Jian Wu, Hongxia Xu Published: 2026-06-25 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.26973v1 Open-set semi-supervised learning aims to leverage unlabeled data that may contain out-of-distribution outliers while maintaining performance on in-distribution classes. Existing methods mainly follow two paradigms: filtering suspicious samples or incorporating unlabeled objectives with soft weighting. We argue that both face a common trade-off: aggressive filtering can discard informative but hard ID samples, whereas utilization can introduce auxiliary gradients that conflict with supervised learning when pseudo labels are wrong. We therefore shift the focus from sample selection to gradient-level control. We propose \textit{Geometric Gradient Rectification} (GGR), a plug-in framework that uses the supervised gradient as an anchor and projects conflicting auxiliary gradients onto an admissible region in gradient space. This makes the applied auxiliary update first-order non-opposing within the rectified coordinate block while preserving orthogonal components that may still carry useful representation signals. We further extend GGR with subspace-aware rectification to stabilize the anchor under noisy mini-batch gradients. Experiments on CIFAR and ImageNet benchmarks show that GGR improves representative OSSL baselines in most settings and yields gains in both closed-set generalization and open-set robustness. Code will be available at https://github.com/JiaheChen2002/GGR. --- ### Computer Vision for MOBA Analytics: A Dataset and Baseline for Visibility Analysis in Dota 2 Authors: Ricardo da Rocha Carvalho, Eloísa Oliveira, Luiz Bernardo Martins Kummer, Emerson Cabrera Paraiso, Rayson Laroca Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26970v1 Introduction: Most Multiplayer Online Battle Arena (MOBA) analytics studies rely on structured data, which does not directly capture what each team could actually see during a match. Objective: This work introduces Dota2-Vis, a video-based dataset, and a baseline pipeline for visibility analysis in professional Dota 2 matches. Methodology: The dataset comprises all 144 matches from The International 2025, recorded from both team perspectives, totaling 288 Full HD videos, together with 2,477 manually annotated minimap images. We evaluate multiple variants of a modern object detector for player-icon detection and use the best-performing model to estimate opponent-visible player presence over time. Results: YOLO11l (large) achieved the best overall performance, reliably identifying player icons even in dense and visually cluttered minimap scenes. The resulting visibility curves reveal player, hero, role, and team-level patterns that complement conventional MOBA analytics, highlighting behavioral differences that are difficult to obtain from structured data alone. The dataset and code are publicly available at https://github.com/RicardoRCarvalho/dota2-vis/. --- ### Einstein World Models Authors: Munachiso Samuel Nwadike, Zangir Iklassov, Ali Mekky, Zayd M. Kawakibi Zuhri, Kentaro Inui Published: 2026-06-25 Categories: cs.AI, cs.CL, cs.CV arXiv: https://arxiv.org/abs/2606.26969v1 Does intelligence require the ability to reason about phenomena beyond direct experience? It is natural to suspect that some complex thought cannot be captured through language alone. However, of particular concern to this work, is whether visualising counterfactual events can complement language as a mechanism for complex thought. We ask whether LLMs can be trained to utilise such visualisation mechanisms, in a way that benefits their reasoning abilities. Motivated by this question, we propose Einstein World Models. EWMs are a blueprint for LLM-based reasoning systems that place visual-temporal rollouts inside the reasoning trace, allowing them to reason in ways that text alone may not support well. In an EWM, the LLM calls a world-module (not to be confused with a world model), to produce short rollouts of scenes under consideration. The returned rollout is treated not as the answer, but as an inspectable hypothesis that can support later reasoning. Einstein World Models extend the capability of LLMs for tool calling (such as web search or code execution), into the domain of visual thought experiments. --- ### RedVox: Safety and Fairness Gaps in Speech Models Across Languages Authors: Beatrice Savoldi, Sara Papi, Wafa Aissa, Matteo Negri, Luisa Bentivogli Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26968v1 Speech-capable models are increasingly deployed in real-world applications across languages. Yet their safety and fairness beyond English settings and under naturalistic conditions remain understudied. We survey safety reporting practices across state-of-the-art speech model releases, finding that only 8% document any multilingual analysis. To address this gap, we introduce RedVox, a multilingual safety and fairness benchmark for audio and speech built on real voices, covering unsafe and unfair stereotypical requests across five languages (English, French, Italian, Spanish, and German). Evaluating eight state-of-the-art models, we find that vulnerabilities persist even under non-adversarial conditions, worsen in non-English languages, and are amplified when the request comes from a spoken input. Finally, by surveying the participants who contributed to RedVox, we document the unique personal and privacy challenges of collecting speech data with human participants, pointing to broader sociotechnical challenges in naturalistic speech safety research. --- ### Look-Before-Move: Narrative-Grounded World Visual Attention in Dynamic 3D Story Worlds Authors: Jiaming Bian, Bingliang Li, Yuehao Wu, Pichao Wang, Zhi Wang, Hailan Ma, Huadong Mo, Zhenhong Sun Published: 2026-06-25 Categories: cs.AI, cs.CV arXiv: https://arxiv.org/abs/2606.26964v2 As embodied AI and world models increasingly operate in dynamic 3D environments, visual perception must move beyond passively interpreting given observations toward actively deciding what to observe. We study this problem through camera planning in dynamic 3D story worlds, where the camera must not only generate smooth motion, but also decide what visual evidence should be acquired before it moves. We formulate this capability as Narrative-Grounded World Visual Attention, where the camera acts as an embodied observer that determines what to observe, how to compose the observation, and how to shift attention over time under narrative intent and physical 3D constraints. To realize this capability, we propose Look-Before-Move, a camera planning framework that separates observation specification from motion execution. It first builds a Semantic Observation Contract to convert directorial intent into executable visual constraints, then performs Monte Carlo Viewpoint Search to find narrative-compliant and geometrically feasible viewpoints, and finally applies Semantic Trajectory Grounding to connect selected viewpoints into continuous, collision-aware, and temporally coherent camera motion. We further construct a dynamic 3D Story World Benchmark based on StoryBlender, covering 50 stories, 457 scenes, and 1585 shots with animated characters, semantic scene configurations, and executable 3D environments. Experiments show that our framework improves subject perception, intent consistency, and trajectory quality over representative baselines, demonstrating the importance of organizing visual attention before generating camera motion. --- ### Term-Centric Hierarchy Induction from Heterogeneous Corpora Authors: Elena Senger, Yuri Campbell, Jan-Peter Bergmann, Rob van der Goot, Barbara Plank Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26963v1 Organizing knowledge from diverse text sources into interpretable hierarchies is crucial for tasks such as policy analysis, innovation monitoring, and exploratory domain mapping. Existing taxonomy induction methods typically rely on document-level representations that capture entire documents rather than the specific domain concepts relevant for knowledge organization, limiting their ability to generalize across heterogeneous sources. We propose a term-centric framework for inducing hierarchical taxonomies from heterogeneous corpora that scales to massive document collections. Our approach maps documents from diverse sources into a shared representation space using automatic term extraction, enabling robust cross-source alignment. Based on these representations, we construct interpretable hierarchies that integrate domain priors with datadriven clustering. Experiments on a novel English and German multi-source benchmark of over one million documents demonstrate that our method improves cross-source coherence and hierarchy quality over text- and summarybased baselines. A case study on German regional innovation analysis further demonstrates its practical utility for technology landscape mapping. --- ### Scaling Multi-Reference Image Generation with Dynamic Reward Optimization Authors: Wenwang Huang, Yusen Fu, Junjie Wang, Mengfei Huang, Yulin Li, Gan Liu, Jing Cai, Yancheng He, Zhuotao Tian Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26947v1 While personalized image generation has achieved remarkable progress, multi-reference image generation (MRIG) remains a challenging task. Most existing benchmarks fail to adequately evaluate complex MRIG scenarios, hindering further progress in this area. To better assess model performance on complex MRIG tasks, we introduce OmniRef-Bench, a benchmark that covers complex combinations of reference image types and a large number of reference images. Evaluations on OmniRef-Bench show that mainstream open-source models struggle in complex MRIG scenarios, and their performance deteriorates significantly as the number of mixed-type reference images increases. To address this issue, we propose DyRef, a two-stage training framework. In the first stage, supervised fine-tuning equips the model with the basic capability to handle complex MRIG tasks. In the second stage, we introduce Difficulty-aware Advantage Reweighting (DAR) and Discriminative Reward Scaling (DRS). DAR dynamically adjusts the optimization objective to improve performance when handling a large number of mixed-type reference images. DRS enlarges intra-group reward differences for more effective policy optimization. Experiments demonstrate that DyRef significantly improves the performance of open-source models on OmniRef-Bench and single-image editing benchmarks, demonstrating the effectiveness and generalization capability of our approach. --- ### TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment Authors: Shuchao Duan, Alan Whone, Hossein Rahmani, Jun Liu, Majid Mirmehdi Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26942v1 Existing facial expression quality assessment (FEQA) methods typically produce only a severity score, without explicitly communicating the observable facial motion evidence that supports the prediction. This limits interpretability and makes it difficult to inspect the basis of model outputs in Parkinson's disease assessment. To address this gap, we propose TraMP-LLaMA, a unified multimodal framework that jointly predicts severity scores and generates structured textual reports from facial motion cues. The framework integrates RGB appearance and landmark trajectory cues, and adopts a decoupled instruction-tuning strategy to reduce task interference between severity prediction and language generation. To support this task, we further extend the PFED5 dataset with expert-guided textual motion descriptions and construct PFED5-plus. Experiments on PFED5-plus show that TraMP-LLaMA outperforms competitive video-language baselines in report generation and achieves the best severity prediction performance among the compared methods under joint multi-expression training, improving Spearman's rank correlation by at least 4.39 percent over all competing methods. The text annotations and code are available at https://github.com/shuchaoduan/TraMP-LLaMA. --- ### Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE Authors: Haoyou Deng, Keyu Yan, Chaojie Mao, Xiang Wang, Yu Liu, Changxin Gao, Nong Sang Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26938v1 Mixture-of-Experts (MoE) architectures have emerged as a powerful paradigm for scaling diffusion models in visual generation. Recent advancements have focused on adaptively allocating computational resources across diverse tokens to improve efficiency and performance. However, we identify a routing assignment problem in existing diffusion MoE frameworks: the router fails to accurately allocate more computational resources to salient tokens. Our analysis attributes this failure to the router's reliance on noise-corrupted latent features throughout the denoising process. Such stochastic noise obscures the critical structural and textural information, thereby preventing the router from effectively distinguishing salient tokens. To address this, we propose SharpMoE, a post-training framework with a saliency-harnessing accurate routing mechanism, which utilizes clean latent features as a noise-free guidance signal for routing. By bypassing the noise-distorted inputs, SharpMoE provides the router with clear saliency guidance, enabling the identification of salient tokens even in high-noise stages. Furthermore, we introduce a trajectory routing loss to constrain the compute allocation throughout the multi-step denoising trajectory, ensuring precise resource allocation along the generation rollout. Extensive experiments demonstrate that SharpMoE serves as a versatile, plug-and-play solution that further enhances the pretrained, converged MoE models, achieving state-of-the-art performance in visual generation. --- ### Jailbreaking for the Average Jane: Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries Authors: Prarabdh Shukla, Ritik, Suhas Rao, Arpit Agarwal, Arjun Bhagoji Published: 2026-06-25 Categories: cs.CR, cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.26936v1 With a profusion of jailbreaks for LLMs now widely known, a growing concern is that non-expert malicious actors ("the average Jane") could elicit actionable responses to malicious requests. In this work, we examine whether this concern is justified. A non-expert malicious actor requires two ingredients for a successful attack: a powerful jailbreak for their target model, acting on an effective malicious query. For the former, we propose a novel attack strategy based on the multi-armed bandit framework. This allows efficient online learning of the optimal jailbreak from a large choice set via noisy exploration on a small number of queries, with subsequent application of the learnt policy on an exploitation set. For the latter, we curate $\mathrm{FrankensteinBench}$, a safety benchmark of $11,279$ malicious queries drawn from manual curation over $7$ existing benchmarks, along with automated enhancement and generation. Each query is categorized as simple or complex by the technical expertise required to craft it. Our findings confirm the concern. Our bandit-based attack achieves success rates as high as $97\%$ on average over $15$ SoTA open-weight LLMs. Moreover, adding complexity to queries raises the attack success rate by up to $26\%$ on average across models -- making it an effective, automatable prompting strategy. --- ### Where Do CoT Training Gains Land in LLM based Agents? Authors: Jingyu Liu, Zhiwen Wang, Yuxin Jing, Huanyu Zhou, Yong Liu Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26935v1 Chain-of-thought (CoT) reasoning is widely used in language-model agents, but prior work has shown that verbalized CoT is not always faithful and may instead reflect post-hoc reasoning, which means the model already knows the answer before reasoning. We therefore ask what CoT training is actually improving: is the model getting better at changing its action through generated reasoning, or is it getting better at predicting the action directly from the prompt? We study this question by comparing \emph{prompt actions} (predicting action without CoT) with CoT actions (predicting action with CoT). Across checkpoints, prompt-action quality improves substantially. While interacting with the environment, the relative advantage of CoT actions over prompt actions remains similar, showing that CoT training does not widen the advantage of CoT reasoning, and it helps to improve the quality of prompt actions. We further find that later checkpoints are less likely to revise the action in response to CoT, suggesting greater reliance on the prompt. Motivated by these patterns, we selectively mask action-token supervision on a fraction of training examples. This intervention improves out-of-domain generalization. --- ### Chai: Agentic Discovery of Cryptographic Misuse Vulnerabilities Authors: Corban Villa, Sohee Kim, Austin Chu, Alon Shakevsky, Raluca Ada Popa Published: 2026-06-25 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2606.26933v1 AI-assisted vulnerability discovery has proven effective for bug classes like memory safety, where instrumentation confirms memory violations and efficiently filters false positives. Many dangerous vulnerability classes, such as cryptographic misuse, however, lack any comparable instrumentation. In this work, we present Chai, an AI-based system that discovers and validates cryptographic misuse vulnerabilities through naturally occurring signals. To achieve this, Chai rethinks the classical technique of differential testing by leveraging AI to 1) improve precision for detecting real security issues in libraries, and 2) repurpose commonly overlooked discrepancies as leads for tangible vulnerabilities in downstream applications. In doing so, Chai inverts the prevailing paradigm of AI vulnerability discovery: instead of auditing one codebase for many flaws, it catalogs flaws at the library level and propagates them across a cryptographic dependency graph, delivering compounding efficiency gains. We evaluate Chai across X.509, JWT, and SAML libraries. Chai discovered a previously unknown critical vulnerability in an SSL library that powers billions of devices, along with security bugs in one library behind a major web browser and another in major Linux distributions. In total, these techniques surfaced over 100 vulnerabilities. --- ### PortraitGen: Exemplar-Driven GRPO with Dual-Reward Guidance for Photorealistic Portrait Generation Authors: Xiaomin Li, Qian Liang, Yinan Li, Ying Zhang, Chen Li, Jing Lyu, Huchuan Lu, Xu Jia Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26930v1 Reinforcement Learning like Group Relative Policy Optimization (GRPO) has significantly advanced text-to-image post-training. However, current methods often favor superficial aesthetics, such as over-saturated colors, leaving critical flaws like AI artifacts and biological implausibilities unresolved. We attribute these limitations to two primary factors: (1) The absence of real images during post-training confines GRPO sampling to the original distribution, failing to break inherent generative boundaries; (2) the optimization process lacks specific rewards targeting fine-grained artifacts like overly oily skin and other AI artifacts. To address this, we propose PortraitGen, a novel framework tailored for photorealistic portrait generation. First, we break inherent generative boundaries by directly introducing real images into the GRPO sampling groups, where image inversion is employed to obtain their transition probabilities and latents. Second, to explicitly steer the model toward photorealism, we introduce a complementary dual-reward mechanism: OmniReward for general quality and AI-Portrait for human-centric fidelity. Furthermore, we curate PortraitBench, a comprehensive portrait-centric benchmark. Extensive experiments demonstrate that PortraitGen significantly outperforms existing baselines, effectively suppressing AI artifacts and achieving unprecedented photorealism. --- ### A Deterministic Control Plane for LLM Coding Agents Authors: Padmaraj Madatha Published: 2026-06-25 Categories: cs.SE, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2606.26924v1 LLM coding harnesses grant agents broad file and shell access, yet the configuration layer that steers them -- rules files, agent definitions, IDE-specific markdown -- is largely unmanaged. A prevalence study of 10,008 public GitHub repositories (n=6,145 agent config files) finds that agent configurations propagate as undeclared shared components: 10.1% of tracked paths are SHA-256 exact duplicates across independent repositories (fork-adjusted, threshold-independent), with 75.5% of clone pairs crossing organisational boundaries. Two further patterns are indicative: configurations are rarely revised (58% single-commit; 0.4 vs 0.6 commits/month age-normalised against CI/CD workflows), and rarely declare permission boundaries (<1% of agent configs vs 33% of Actions workflows, n=31 true positives). We propose a deterministic control plane above the harness that maps one-to-one to these gaps. Rel(AI)Build treats agent definitions as a managed supply chain (SHA-256 content addressing, HMAC-stamped lockfiles, hash-chained audit logs); enforces tiered permissions and attack-derived blocklists before LLM invocation; gates feature work through a phase state machine with requirement-to-file-to-test traceability; compiles a single canonical definition to seven IDE targets; and detects prompt drift via Jaccard similarity. Conformance tests on injected violations confirm each mechanism enforces its stated invariant; developer outcomes remain future work. Governance of this layer must be deterministic and tool-agnostic -- not delegated to further LLM orchestration. --- ### GAVEL: Grounded Caption Error Verification and Localization Authors: Zixian Gao, Atsushi Hashimoto, Kuniaki Saito Published: 2026-06-25 Categories: cs.CL arXiv: https://arxiv.org/abs/2606.26923v1 Vision-language models (VLMs) often produce hallucinated or inconsistent outputs, where text and images are not properly aligned. Addressing this issue requires not only detecting misalignment but also explaining the discrepancy and localizing its visual evidence. We introduce GAVEL (Grounded Caption Error Verification and Localization), a task that jointly addresses verification, explanation, and localization for image-text pairs. To support systematic evaluation, we also present a corresponding dataset and benchmark. We further train a supervised baseline on the human-annotated training split to assess whether GAVEL provides learnable supervision for these abilities. Experiments show that even strong closed-source models struggle on GAVEL, while the supervised baseline yields consistent improvements across grounding and explanation metrics. --- ### Risk-Aware Selective Multimodal Driver Monitoring with Driver-State World Modeling Authors: Daosheng Qiu, Haozhuang Chi, Hao Su, Shu Long, Xinyue Miao, Yongle Dong, Wei Zhang Published: 2026-06-25 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2606.26922v1 Continuous driver monitoring in automated vehicles requires low-latency inference while avoiding unsafe decisions under uncertain driver states. Large vision-language models provide broad multimodal priors, but their latency and limited reliability in this setting make them unsuitable as always-on in-cabin monitors. We propose a cost-aware selective inference framework for deployable multimodal driver monitoring. The core system is a lightweight RGB-physiological student that combines in-cabin visual observations with window-level HR/EDA signals, and a learned gate that decides when to accept the fast prediction or abstain for safety intervention. Additional controls show that the learned scores contain sample-level information beyond scenario priors, while exact physiological synchronization remains a limitation. To incorporate predictive evidence, we further study a compact driver-state world modeling module that rolls out latent driver-state features and estimates future fast-model errors and counterfactual system-level action costs. On scenario-induced driver-demand recognition, the RGB-physiological student improves over RGB-only and physiology-only baselines, reaching 0.7440 Macro-F1 and 0.9099 balanced accuracy with 11.39M parameters and 3.08ms inference latency. Cost-aware selective inference reduces unsafe false negatives from 17.37% under always-fast inference to approximately 5% across seeds, while maintaining deployment-level latency. While driver-state world modeling offers valuable predictive signals, worst-group evaluations highlight persistent operating-point calibration drift. Ultimately, reliable edge driver monitoring requires advancing not only perception backbones, but also risk-aware selective control and group-robust calibration. --- ### Diagnosing Task Insensitivity in Language Agents Authors: Jingyu Liu, Xiaopeng Wu, Kehan Chen, Chuan Yu, Yong Liu Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26918v1 Large language models can serve as capable long-horizon agents, but their out-of-distribution (OOD) generalization remains weak. We identify a key source of this failure as task insensitivity: when faced with similar but distinct tasks, models might apply patterns learned during training and fail to solve the task at hand. We show that models often continue with actions aligned with the original task even when the instruction is semantically corrupted and cannot be directly answered. We further find that, when we replace the task description in a trained prompt with another similar but distinct task, the model may still output the same action. This behavior is accompanied by a consistent training-time attention drift away from task tokens and toward local observations, suggesting an optimization bias toward shortcuts. To mitigate this problem, we propose Task-Perturbed NLL Optimization, a lightweight contrastive regularizer that explicitly encourages action dependence on the task instruction. Extensive evaluations show that our intervention improves task sensitivity and OOD generalization while preserving more stable attention to task tokens. --- ### GEOALIGN: Geometric Rollout Curation for Robust LLM Reinforcement Learning Authors: Ting Zhou, Zhenqing Ling, Yiyang Zhao, Ying Shen, Daoyuan Chen Published: 2026-06-25 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2606.26917v1 Online reinforcement learning is widely used to align large language models (LLMs) with reward signals, yet training can be unstable under noisy or misspecified rewards. We identify a failure mode we call directional inconsistency: within a batch, a small set of high-reward rollouts induces representation-space preference directions that sharply disagree with the batch majority, resulting in high-variance and destabilizing updates. We propose geoalign, a lightweight plug-in for rollout curation in iterative policy optimization. Geoalign (i) forms within-prompt preference pairs, (ii) learns an online projector on per-rollout hidden states to concentrate reward-ordered displacement directions, and (iii) detects directionally inconsistent rollouts via their angular deviation from a batch consensus prototype and rectifies them with within-prompt stable alternatives. Geoalign is forward-pass only and adds negligible overhead. Across dialogue alignment with a learned reward model and mathematical reasoning with binary verified rewards, Geoalign improves final performance and reduces training oscillation, outperforming PF-PPO, PAR, PODS, and Seed-GRPO. These results suggest latent directional consensus as an effective reliability signal for online LLM RL. --- ### PhysRAG: Enhancing Physics-Awareness in Video Generation via Retrieval-Augmented Generation Authors: Kexu Cheng, Zicheng Liu, Mingju Gao, Chunhe Song, Hao Tang Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26916v1 Developing physically aware video generation models remains a significant challenge due to the difficulty in capturing diverse physical phenomena, such as thermal dynamics, mechanics, and optics. In this work, we introduce PhysRAG, a novel pipeline that enhances physical awareness in video generation through Retrieval-Augmented Generation (RAG). To address the issue of limited high-quality data, we design a two-stage data filtering pipeline based on the WISA-80K dataset, resulting in a curated set of 7K high-quality videos for training. Furthermore, we construct a physical video database and develop a mechanism to inject physical knowledge into a video diffusion model using learnable queries. Our method achieves state-of-the-art performance in both visual quality and physical rule compliance, surpassing existing models in benchmarks such as PhyGenBench and VBench. We conduct extensive ablation studies to validate the effectiveness of our key components, including the data filtering pipeline, RAG mechanism, and method for physical information extraction. To facilitate future research, our code, data, and models are prepared for release at https://github.com/sediment1024/PhysRAG. --- ### Neural Texture Compression using Hypernetworks Authors: Belcour Laurent Published: 2026-06-25 Categories: cs.GR, cs.CV arXiv: https://arxiv.org/abs/2606.26913v1 Recent work on neural texture compression has demonstrated that it is possible to learn small, per-material texture representations (composed of latent textures and a small Multi-Layer Perceptron decoder) that can be decoded in real-time during shading to reproduce the input to a physically based shading model. However, existing methods require performing gradient-descent optimization per material for a given MLP and latent configuration. In this work, we train a single hypernetwork that outputs both the latent features and the MLP's weights and biases. Though the solution space is high-dimensional, this approach produces results comparable in quality to the current reference neural texture compressors. We further extend this approach to infer multiple decoders at once or even produce decoders that learn super-resolution. --- ### Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation Authors: Zekai Zhang, Jiahao Li, Jie Zhang, Kaiyuan Gao, Kun Yan, Lihan Jiang, Ningyuan Tang, Shengming Yin, Tianhe Wu, Xiaoyue Chen, Xiao Xu, Yan Shu et al. Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26907v2 While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation context for T2I models. To bridge this gap, we propose Qwen-Image-Agent, a unified agentic framework that integrates plan, reason, search, memory and feedback in a context-centric manner. Qwen-Image-Agent treats user input as partial context and progressively constructs the generation context through Context-Aware Planning and Context Grounding. Specifically, Context-Aware Planning identifies missing context and plans how it should be acquired and used, while Context Grounding gathers this context from reason, search, memory, and feedback. To evaluate agentic image generation, we further introduce Image Agent Bench (IA-Bench), a benchmark covering four core image agent capabilities: Plan, Reason, Search, and Memory. Experiments on IA-Bench, Mindbench and WISE-Verified show that Qwen-Image-Agent outperforms strong baselines and achieves state-of-the-art performance. --- ### Confidence-Aware Tool Orchestration for Robust Video Understanding Authors: Yangfan He, Yujin Choi, Jaehong Yoon Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26904v1 Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world embodied benchmarks, while remaining unaware that their visual evidence has been degraded. To address this challenge, we propose Robust-TO, an agentic video understanding framework that explicitly integrates per-frame trustworthiness into every stage of reasoning. Robust-TO organizes heterogeneous visual perception tools under a unified evidence interface. Each tool receives a sub-query derived from the original question and a set of trustworthy frames selected by the reliability-relevance score. It returns evidence in a shared format: a concrete prediction (e.g., a bounding box, motion trajectory, recognized text, or action label), temporal grounding, and a calibrated reliability score. During reasoning, these calibrated scores guide evidence weighting in a three-tier synthesis process (high/medium/low) and define a confidence-cost GRPO reward that jointly optimizes correctness, evidence reliability, and efficiency. On two video reasoning benchmarks spanning eight tasks, Robust-TO achieves 56.4% average accuracy on clean inputs, surpassing the strongest open-source baseline by 10.6%p and outperforming Gemini-2.5-Pro (46.2%). Under five realistic corruption types, Robust-TO maintains 54.3% average accuracy, 5.8%p above the strongest open-source baseline, while exhibiting the smallest clean-to-corrupted accuracy drop among all compared methods. --- ### Learning to Recover Task Experts from a Multi-Task Merged Model Authors: Jinwook Jung, Taegyu Kim, Kumju Jo, Sungyong Baik Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26902v2 Multi-task model merging aims to consolidate several task-specific experts into a unified model, yet static merging consistently suffers from parameter interference. While dynamic merging models aim to bridge this gap, many works rely on the costly storage and loading of redundant expert components at inference. In this work, from the perspective of task expert, we view parameter interference as parameter perturbation introduced to each expert during merging process. We show that such parameter perturbations can be modeled as affine transformation, which can be approximated as additive offsets. Motivated by these, we propose Recover Task eXpert (ReTeX), a framework that predicts those offsets, in order to undo parameter interference and recover task-expert performance from a single merged checkpoint. To recover the appropriate expert when task identity is unknown, we introduce a router-free task identifier based on SVD subspace signatures computed offline before inference. At inference, the identifier selects the task whose subspace yields the smallest projection residual for a given input. As a result, ReTeX recovers over 95% of individual-expert performance in both vision and NLP domains, while significantly improving generalization to unseen tasks. Crucially, we also show that the parameter offset prediction leads to emergent adaptive interpolation of expert knowledge for out-of-distribution (OOD) tasks. ReTeX adaptively interpolates seen expert knowledge to handle unseen tasks. Our code is available at https://github.com/BAIKLAB/ReTeX --- ### Learning to Recover Task Experts from a Multi-Task Merged Model Authors: Jinwook Jung, Taegyu Kim, Kumju Jo, Sungyong Baik Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26902v1 Multi-task model merging aims to consolidate several task-specific experts into a unified model, yet static merging consistently suffers from parameter interference. While dynamic merging models aim to bridge this gap, many works rely on the costly storage and loading of redundant expert components at inference. In this work, from the perspective of task expert, we view parameter interference as parameter perturbation introduced to each expert during merging process. We show that such parameter perturbations can be modeled as affine transformation, which can be approximated as additive offsets. Motivated by these, we propose Recover Task eXpert (ReTeX), a framework that predicts those offsets, in order to undo parameter interference and recover task-expert performance from a single merged checkpoint. To recover the appropriate expert when task identity is unknown, we introduce a router-free task identifier based on SVD subspace signatures computed offline before inference. At inference, the identifier selects the task whose subspace yields the smallest projection residual for a given input. As a result, ReTeX recovers over 95% of individual-expert performance in both vision and NLP domains, while significantly improving generalization to unseen tasks. Crucially, we also show that the parameter offset prediction leads to emergent adaptive interpolation of expert knowledge for out-of-distribution (OOD) tasks. ReTeX adaptively interpolates seen expert knowledge to handle unseen tasks. Our code is available at https://github.com/BAIKLAB/ReTeX --- ### SamaVaani: Auditing and Debiasing Multilingual Clinical ASR for Indian Languages Authors: Subham Kumar, Prakrithi Shivaprakash, Abhishek Manoharan, Astut Kurariya, Diptadhi Mukherjee, Prabhat Chand, Pratima Murthy, Koustav Rudra, Lekhansh Shukla, Animesh Mukherjee Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.26901v1 Automatic Speech Recognition (ASR) is increasingly used to document clinical encounters, yet its reliability in multilingual and demographically diverse Indian healthcare context remains largely unknown. In this study, we first conduct the systematic audit of ASR performance on real-world psychiatric interview data spanning Kannada, Hindi and Indian English, comparing eight state-of-the-art models including IndicWhisper, WhisperLargeV3, Sarvam, GoogleS2T, Gemma3n, OmniLingual, Vaani, and Gemini. Our results reveal substantial variability across models and languages, with some systems performing competitively in Indian English but failing in regional speech. We further fine-tune two of the best performing opensource models, i.e., Gemma3n and OmniLingual, using various methods. With this, we uncover systematic performance gaps tied to speaker role and gender, raising concerns about equitable deployment in clinical settings, which are further mitigated by fairness-aware fine-tuning. To this end, we propose SamaVaani, a unified debiasing technique that simultaneously improves ASR performance and improves fairness across demographic groups. --- ### Generative Retrieval via Diffusion Transformer with Metric-Ordered Sequence Training and Hybrid-Policy Preference Optimization Authors: Chenghao Liu, Yu Zhang, Zhongtao Jiang, Kun Xu, Zhenwei An, Renzhi Wang, Zhao Wang, Jiachen Zhang, Yuxiao Zhang, Kun Xu, Songfang Huang Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26899v1 Embedding-based retrieval ranks items by their similarity to a query in a shared vector space and usually aims to return the highest-scoring items. In many production settings this is not what is wanted: given a seed set that expresses a fine-grained pattern, one needs more items that both satisfy a target attribute and stay within that pattern. We formalize this as pattern-preserving attribute retrieval. The two goals pull against each other: averaging the seeds preserves the pattern but stays in a low-attribute region, while global attribute retrieval drifts to unrelated patterns. We approach the task with continuous generative retrieval, where a model reads a sequence of item embeddings and generates query embeddings for nearest-neighbor search. We propose MO-DiT+HPPO, a staged framework with raw-sequence pretraining, multi-domain metric-ordered continuation pretraining, tail-centroid fine-tuning, and HPPO. Metric-ordered training turns sparse online retrieval labels into in-pattern trajectories ordered from low to high predicted attribute density, teaching one model the metric-improvement direction across domains. HPPO aligns the generated query distribution with the true online objective by labeling a hybrid candidate pool with the online intersection metric and applying reference-anchored preference optimization. A Pareto pair filter keeps only winner pairs that do not lower same-pattern purity, raising the attribute metric without sacrificing the pattern. Across four attribute domains under item- and pattern-holdout protocols, metric-ordered DiT improves the intersection metric over a pretrained generative retriever, and HPPO improves it further, with significant gains on seven of eight domain-split cells and a marginal tie on the hardest split. Metric-predictor validation, order ablations, CPT/SFT comparisons, and a candidate-policy ablation show where the gains come from. --- ### Tractography-Driven Synthetic Data Generation for Fiber Bundle Segmentation in Tracer Histology Authors: Kyriaki-Margarita Bintsi, Sparsh Makharia, Yaël Balbastre, Joselyn Romero Avila, Julia F. Lehman, Suzanne N. Haber, Anastasia Yendiki Published: 2026-06-25 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2606.26898v1 Diffusion MRI (dMRI) tractography enables non-invasive reconstruction of white-matter pathways, but its accuracy is fundamentally limited by indirect, low-resolution measurements of axonal organization. Tracer injection studies in non-human primates provide a gold standard for validating dMRI tractography. This, however, requires time-consuming manual annotation of fiber bundles in histology sections. We propose a synthetic-data augmented framework for automated fiber bundle segmentation in macaque tracer histology. Our approach uses ex vivo dMRI tractography as a generative prior to synthesize 2D image patches for training. This provides us with sufficiently realistic foreground texture, which we compose with backgrounds from blockface photos and diversify via domain randomization. A 2D U-Net is trained on mixed real and synthetic patches. Experiments on held-out brains demonstrate improved generalization across brains and fiber bundle densities compared to training with real data only. Training with synthetic data only leads to poor performance, underscoring the need for real supervision. Overall, our approach achieves performance comparable to the state-of-the-art while requiring 3x less manually annotated data. --- ### Modeling Local, Global, and Cross-Modal Context in Multimodal 3D MRI Authors: Minh Duc Do, Tillmann Rheude, Noel Kronenberg, Roland Eils, Benjamin Wild Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26894v1 Brain MRI poses a fundamental challenge for machine learning: models must learn from high-dimensional 3D data spanning multiple co-registered modalities, despite the limited sample sizes typical of neuroimaging studies relative to the diversity in anatomy, pathology, and acquisition conditions. While multimodal imaging provides complementary information critical for clinical interpretation, effectively integrating these signals remains difficult. We propose Multimodal Intra- and Cross-Context Vision Transformer (MICViT), a 3D vision transformer that explicitly models both modality-specific representations and cross-modal interactions across local and global contexts. Concretely, MICViT combines four attention mechanisms: modality-specific local and global attention for intra-modal feature learning, and cross-modal local and global attention to capture interactions between modalities. We evaluate MICViT on brain age prediction across three heterogeneous datasets (UK Biobank, n=41,404; SOOP, n=1,062; Cam-CAN, n=613) using multiple MRI modalities (e.g. T1, FLAIR, DWI, SWI). MICViT consistently outperforms state-of-the-art CNN and transformer baselines in 3D settings. Notably, it benefits more strongly from multimodal inputs, yielding larger performance gains as additional modalities are incorporated. These results demonstrate that explicitly modeling intra- and cross-modal interactions is key to unlocking the full potential of multimodal brain MRI, highlighting a promising direction for representation learning in neuroimaging. --- ### Asymptotically Optimal Learning for Parametric Prophet Inequalities Authors: Jung-hun Kim, Anna Grebennikova, Vianney Perchet Published: 2026-06-25 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2606.26893v1 We study learning in prophet inequalities with i.i.d. rewards drawn from an exponential-type parametric family with an unknown parameter $θ$, a class that includes exponential, Pareto, and bounded-support power-family distributions. We first characterize the optimal full-information asymptotic competitive ratio for this family. In the unbounded-support case, the limit is $ {\left(θ/({θ-c_+})\right)^{c_+/θ}}/ {Γ(1-c_+/θ)},$ while in the bounded-support case, the limit is $1$. We then propose a confidence-based dynamic-programming policy for online learning. By exploiting the explicit parametric structure, the policy achieves the same optimal asymptotic competitive ratio using only online observations, without external offline samples. We further derive distribution-specific convergence rates for canonical examples. Finally, numerical experiments on synthetic instances illustrate the performance of our algorithm. --- ### Bridging Vision and Language Concepts through Optimal Transport Semantic Flow Authors: Chenyang Zhang, Anqi Dong, Guangming Zhu, Nuoye Xiong, Siyuan Wang, Lin Mei, Liang Zhang Published: 2026-06-25 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2606.26891v1 Concept Bottleneck Models (CBMs) promise transparent reasoning by predicting through human-interpretable concepts, yet their effectiveness fundamentally depends on how well visual and textual representations are aligned or matched. Existing vision-language CBMs often rely on pre-aligned encoders or global cosine similarity, which obscures fine-grained concept localization and fails to reflect true semantic geometry. In this work, we rethink concept alignment as a dynamic cross-modal transport process instead of static projection and propose the Optimal Transport Flow Concept Bottleneck Model (OTF-CBM). It first learns a data-driven semantic cost via Inverse Optimal Transport to measure cross-modal distances, and then performs unbalanced optimal-transport-based flow matching to model semantic transitions between visual patches and textual concepts. With velocity-based concept activation, OTF-CBM captures interpretable geometric relations without ODE integration. Experiments further show that OTF-CBM achieves superior classification accuracy and concept faithfulness, offering a new geometric and dynamical perspective for interpretable cross-modal reasoning. --- ### RIS-Assisted Proactive Handover for Reliable mmWave Wireless Networks Authors: Alaa Adnan, Mohammad Al-Quraan, Ahmed Zoha, M. Majid Butt, Sami Muhaidat, Muhammad Ali Imran, Marco Di Renzo, Lina Mohjazi Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26885v1 Millimeter-wave (mmWave) networks are highly susceptible to line-of-sight (LoS) blockages. Vision-aided wireless communications (VAWC) enable proactive handovers (PHO) to mitigate such blockages; however, PHO becomes challenging when no nearby base station (BS) is available. In such cases, reconfigurable intelligent surfaces (RIS) can be used to restore connectivity. To ensure timely PHO, the RIS configuration time must be taken into account, as the large number of RIS elements can limit responsiveness in time-sensitive scenarios. This work proposes a novel RIS-assisted PHO approach that optimizes the number of allocated RIS elements to balance signal processing complexity and link quality under handover timing constraints, making the RIS-assisted link more energy-efficient. An optimization problem based on particle swarm optimization (PSO) is formulated to determine the optimal end-to-end RIS link setup that runs offline to bypass latency constraints. Results show that reducing the number of RIS elements by 12\% leads to a 10\% decrease in dissipated energy without compromising the signal-to-noise ratio (SNR). Moreover, the RIS-assisted link achieves a 15--30 dB improvement in blocked regions while maintaining accurate PHO timing. --- ### Accelerated sampling using SamAdams variable timesteps and position-adaptive Langevin dynamics Authors: Benedict Leimkuhler, Peter A. Whalley Published: 2026-06-25 Categories: math.NA, cs.LG, stat.CO arXiv: https://arxiv.org/abs/2606.26881v1 We introduce an accelerated Langevin-based sampling method that is based on two complementary devices: \emph{SamAdams} adaptive timestepping, which automatically shrinks the effective integration step in stiff regions of phase space using a relaxed stiffness monitor, and \emph{position-adaptive Langevin} (PAL) dynamics, which concentrates friction along the local force direction while preserving the canonical distribution as the exact invariant measure. The resulting combined scheme (SA-PAL) is implemented in a palindromic integrator which requires only one force evaluation per iteration through suitable organisation of the integration steps and by exploiting the rank-one-plus-scalar structure of the PAL friction tensor. We test the method on various model problems: the Rosenbrock function, a thin entropic channel, the Mueller-Brown potential, and a Bayesian parameterisation problem with a sparsity-inducing shrinkage prior. On the Rosenbrock and Mueller-Brown potentials mixing rates are improved by 1.5-3 times compared to fixed stepsize integration. Efficiency gains of more than an order of magnitude are documented in the other examples. --- ### Heterogeneous Neural Predictivity from Language Models During Naturalistic Comprehension Authors: Xiao Jia Published: 2026-06-25 Categories: cs.CL, cs.LG arXiv: https://arxiv.org/abs/2606.26880v1 Language-model representations provide structured, high-dimensional annotations of naturalistic language stimuli and can serve as informative neural predictors during comprehension. We analyzed locked derived data from Brain Treebank, MEG-MASC, and Podcast ECoG with eight frozen language models, blocked encoding models, and matched temporal, nuisance, and representation-capacity controls. Positive held-out prediction and gains over low-level baselines were widespread in source-level summaries. Across Brain Treebank and Podcast ECoG, 67 of 432 evaluable rows met a controlled predictive-only criterion, and model-side feature ablations changed prediction scores in most evaluable source rows. Brain-derived, timing-linked, acoustic, and implanted-signal controls confirmed component-level sensitivity of the analysis pipeline. These findings show that language-model-derived quantities can annotate neural activity during natural speech and text comprehension. Participant-level matched-control advantages were localized rather than uniform, response-profile and feature-specificity contrasts bounded representational or computational interpretations, and complete co-indexed integrated interpretation will require future jointly indexed coverage. Together, the analyses identify language-model features as useful neural predictors and separate predictive usefulness from claims about shared neural organization or language-processing computations. --- ### A Pipeline for Generating Longitudinal Synthetic Clinical Notes Using Large Language Models Authors: William Poulett, Alice Waterhouse, Ben Wallace, Scarlett Kynoch, Amaia Imaz Blanco, Michael Spence, Jonathan Pearson Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26879v2 Synthetic data is increasingly used to enable the development and evaluation of AI systems in domains where access to real-world data is restricted. In healthcare, clinical documentation presents particular challenges due to its sensitivity. This work introduces a synthetic clinical notes pipeline and dataset designed to support the development of clinical AI tools while avoiding the privacy risks associated with real patient data. The dataset is generated using a modular pipeline that combines structured patient generation, semi-structured patient journey simulation, and unstructured clinical note generation using large language models. The pipeline is designed to prioritise internal consistency across longitudinal patient records, while also capturing variation in writing style, note structure, and clinical detail. Additional mechanisms, including LLM-based validation and augmentation steps, are used to improve faithfulness, realism, and diversity of the generated notes. We release a dataset of 70 synthetic patients, each associated with 20-50 clinical notes spanning a full hospital journey. The dataset is provided at multiple levels of validation, enabling users to balance realism and scalability depending on their use case. This dataset supports the development, testing, and evaluation of clinical AI systems, including summarisation tools, coding models, and decision support systems, without reliance on real patient data. --- ### Information-Aware KV Cache Compression for Long Reasoning Authors: Jushi Kai, Zhuiri Xiao, Alexandra Birch, Zhouhan Lin Published: 2026-06-25 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2606.26875v1 Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively captures contextual relevance, it overlooks complementary information-theoretic signals related to predictive uncertainty and token informativeness. In this paper, we revisit token importance from a forward-looking perspective and introduce \textit{Forward Influence}, a metric that measures how compressed tokens affect future contexts. Our analysis reveals that tokens selected by attention scores mainly influence nearby contexts, whereas tokens associated with high predictive uncertainty exhibit substantially stronger influence on distant future contexts. Based on the observation, we propose \textbf{InfoKV}, an entropy-aware KV cache compression framework that incorporates information-theoretic signals. It combines token-level predictive uncertainty with layer-wise representation evolution and integrates the resulting entropy scores with attention scores during reasoning. Experiments on long-context reasoning benchmarks with Llama-3.1, Llama-3.2, and DeepSeek-R1 demonstrate that InfoKV consistently outperforms existing attention-based KV compression methods in both long prefilling and decoding scenarios. --- ### TAVR-VLM: Risk-Conditioned Causal Grounding for Hallucination-Resistant Report Generation Authors: Zhixiang Lu, Xiwei Liu, Sifan Song, Changkai Ji, Anh Nguyen, Jionglong Su, Imran Razzak, Jinfeng Wang Published: 2026-06-25 Categories: cs.AI arXiv: https://arxiv.org/abs/2606.26874v1 Transcatheter Aortic Valve Replacement (TAVR) planning requires meticulous multimodal reasoning. However, adapting Multimodal Large Language Models (MLLMs) to this high-stakes domain is severely impeded by diagnostic hallucinations, where generated text lacks anatomical grounding. To address this, TAVR-VLM is introduced: a novel framework featuring Risk-Conditioned Causal Grounding Attention (R-CGA) that instantiates a model-internal ``Risk $\rightarrow$ Region $\rightarrow$ Word'' structural grounding pathway. R-CGA compresses multimodal inputs into a causal risk bottleneck, purifying dense visual features into a global risk mask. During autoregressive generation, a support-projected causal consistency objective constrains token-level grounding within the risk-defined support mask. Evaluated on $\text{M}^3\text{TAVR}$, a comprehensive 1,482-patient cohort, TAVR-VLM establishes a new state-of-the-art. It achieves an AUROC of 0.896, boosts CIDEr to 0.936, and drastically reduces the hallucination rate to 8.1\%, thereby improving interpretability for evidence-based surgical AI. --- ### Scalable Message-Passing Quantum Graph Neural Networks in the Weisfeiler-Leman Hierarchy Authors: Snehal Raj, Brian Coyle, Léo Monbroussou, André J. Ferreira-Martins, Renato M. S. Farias, Elham Kashefi Published: 2026-06-25 Categories: quant-ph, cs.LG arXiv: https://arxiv.org/abs/2606.26873v1 Graphs provide a natural language for relational data in chemistry, biology and optimisation. Graph neural networks (GNNs) have driven much of the recent progress in learning from such data through message passing, a single primitive that generalises convolution and attention. Quantum counterparts have been proposed, but with limited connection to message passing and few guarantees on performance or scalability. More broadly, the trainability of variational quantum circuits is a recognised bottleneck for their wide applicability, and pre-training has emerged as one way to address it. Yet for a quantum model to be useful, it must offer expressivity guarantees along with demonstrable scalability. Here we show how a quantum graph neural network can be built to perform message passing, to be permutation equivariant, and to sit at a chosen level of the Weisfeiler-Leman hierarchy, the standard measure of how finely a model can tell graphs apart. We show that, as for classical GNNs, the training can be done first on small graph instances, allowing for a pre-training that can mitigate usual training issues, and its output can be read out at a cost that stays low as the graph grows. We validate the framework in large-scale simulations of up to 56 qubits across three datasets, on synthetic graphs that ordinary message passing cannot separate, on molecular property prediction, and on the travelling salesperson problem. Our framework opens a path for near-term quantum algorithms with theoretical guarantees and practical scalability, bringing the principles of graph learning into quantum circuit design. --- ### SpatialFlow-GRPO: Where Spatial Credit Drives Image Editing Authors: Yankai Yang, Yancheng Long, Wei Chen, Xingyu Lu, Hongyang Wei, Bin Wen, Fan Yang, Tingting Gao, Han Li, Shuo Yang Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26872v2 Recent online reinforcement learning has substantially improved image editing quality. However, existing Flow-GRPO-style methods usually rely on a single whole-image reward, which makes fine-grained editing optimization difficult. We observe that a key obstacle in image editing is this spatial uniformity assumption: a whole-image reward cannot distinguish how different spatial regions contribute to image quality. To address this issue, we propose SpatialFlow-GRPO, a training framework that introduces spatially fine-grained reward feedback. The framework converts region-aware rewards into semantic-region-level optimization signals and aligns region advantages with the corresponding latent positions during policy updates. We also train a region-aware reward model, SFReward, construct SFReward-14K with region-annotated editing samples, and introduce MultiEditBench to evaluate multi-region editing ability. On OmniGen2 and FLUX.2-klein-4B, SpatialFlow-GRPO outperforms Flow-GRPO on GEdit-Bench, ImgEdit-Bench, and MultiEditBench. The results show that SpatialFlow-GRPO converts local feedback into spatially aligned update signals and improves editing quality. --- ### Fortress and Gatekeeper: Theorizing Transitive Trust in Third-Party Cybersecurity Risk Governance Authors: Yijun Chen, Misita Anwar Published: 2026-06-25 Categories: cs.CR, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2606.26866v1 Third-party vendors, such as analytics platforms, cloud services, identity providers, and software suppliers, are increasingly embedded in digital service delivery. While these arrangements enable scale and specialization, they also move customer data and security-relevant practices into environments that customers rarely see, select, or evaluate. This paper examines this problem through a document analysis of the November 2025 OpenAI-Mixpanel security incident. The incident serves as an illustrative case for showing how a security event in a vendor environment can become a governance and accountability problem for the focal organization that maintains the customer relationship. Drawing on organizational trust research and agency theory, the paper argues that third-party cybersecurity risk is both a trust relationship and a delegation problem. Customers trust the visible service provider, while the provider relies on vendors whose security practices are only partially visible and controllable. The paper develops the concept of transitive trust, where customer trust in a digital service depends on the security practices of vendors authorized by that service provider. It then presents the Fortress and Gatekeeper framework, which explains cybersecurity governance boundaries through trust and data flows rather than formal organizational ownership alone. The analysis develops four propositions concerning vendor integration, metadata exposure, vendor assurance, and data proliferation. The paper contributes to cybersecurity governance scholarship by explaining how delegated data processing creates customer-facing accountability and by identifying implications for vendor tiering, data classification, contractual design, continuous assurance, and data minimization. --- ### Rolling Shutter Relative Pose Estimation Made Practical Authors: Daniel Barath Published: 2026-06-25 Categories: cs.CV arXiv: https://arxiv.org/abs/2606.26863v1 Rolling shutter (RS) cameras equip virtually all consumer devices, yet RS-aware relative pose estimation has remained impractical: the state-of-the-art solver requires a minimum of 20 point correspondences, making RANSAC-based robust estimation prohibitively expensive due to the exponential dependence of the iteration count on the sample size. We make RS relative pose estimation practical by introducing affine correspondences (ACs) into the RS two-view geometry. We derive novel \emph{RS-corrected affine constraints} that account for the coupling between point perturbations and the row-dependent essential matrix, providing two equations per correspondence beyond the standard epipolar constraint. Building on these constraints, we develop a linearized algebraic solver that estimates pose and RS motion from only 7 ACs. The solver exploits the physical smallness of RS parameters to linearize the constraints, eliminates the 12 RS unknowns via null-space projection, and solves the remaining degree-20 system via action matrices in 1.2\,ms. On the TUM RS benchmark, our method achieves the best pose and RS parameter accuracy among all tested methods and, uniquely among RS solvers, provides accurate translational velocity estimates -- which are poorly conditioned from point correspondences alone due to a $\vec{v}$-$\vec{t}$ coupling. On the global-shutter EuRoC MAV dataset, the solver achieves comparable accuracy to the standard 5-point algorithm, demonstrating that it generalizes well to the GS setting. Code is at https://github.com/danini/rolling_shutter_made_practical. ---