# Research Papers — 2026-07 · bundle page 9 of 222 > Full-text markdown of papers #401–450 of 11075 published in 2026-07, newest first. AI-readable mirrors of the full 11075-paper 2026-07 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-07/page/9 [← Previous bundle](https://aipod.dev/papers/archive/2026-07/page/8.md) [Next bundle →](https://aipod.dev/papers/archive/2026-07/page/10.md) ### Classification of COVID-19 cases from chest CT volumes using hybrid model of 3D CNN and 3D MLP-Mixer Authors: Masahiro Oda, Tong Zheng, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28978v1 This paper proposes an automated classification method of COVID-19 chest CT volumes using improved 3D MLP-Mixer. Novel coronavirus disease 2019 (COVID-19) spreads over the world, causing a large number of infected patients and deaths. Sudden increase in the number of COVID-19 patients causes a manpower shortage in medical institutions. Computer-aided diagnosis (CAD) system provides quick and quantitative diagnosis results. CAD system for COVID-19 enables efficient diagnosis workflow and contributes to reduce such manpower shortage. In image-based diagnosis of viral pneumonia cases including COVID-19, both local and global image features are important because viral pneumonia cause many ground glass opacities and consolidations in large areas in the lung. This paper proposes an automated classification method of chest CT volumes for COVID-19 diagnosis assistance. MLP-Mixer is a recent method of image classification using Vision Transformer-like architecture. It performs classification using both local and global image features. To classify 3D CT volumes, we developed a hybrid classification model that consists of both a 3D convolutional neural network (CNN) and a 3D version of the MLP-Mixer. Classification accuracy of the proposed method was evaluated using a dataset that contains 1205 CT volumes and obtained 79.5% of classification accuracy. The accuracy was higher than that of conventional 3D CNN models consists of 3D CNN layers and simple MLP layers. --- ### Extrapolating the emergence of Hamiltonian chaos with random-feature Hamiltonian neural networks Authors: Jaesung Choi Published: 2026-07-31 Categories: nlin.CD, cs.LG, physics.comp-ph arXiv: https://arxiv.org/abs/2607.28977v1 Machine learning of Hamiltonian dynamics has driven growing interest in Hamiltonian neural networks (HNNs), which encode Hamilton's equations of motion into the learning architecture. Despite this progress, it remains unknown whether such networks can predict dynamical regimes absent from their training data, in particular the broad chaotic sea that emerges beyond the observed parameter interval. We address this question using a parameter-aware random-feature Hamiltonian neural network (RF-HNN). Trained using data from only a small number of control-parameter values at which invariant tori dominate, the RF-HNN predicts autonomous long-time dynamics at unseen parameter values where mixed phase space develops and chaotic regions expand, with no data from that regime used in training or model selection. The method is demonstrated across four two-degree-of-freedom Hamiltonian families, including the Hénon-Heiles system. Using Poincaré-section geometry and finite-time Lyapunov exponents, we show that the RF-HNN reproduces the breakup of regular structures and the emergence and growth of chaotic regions, whereas conventionally trained HNNs with the same Hamiltonian structure remain too regular. These results show that what decides parameter extrapolation is not Hamiltonian structure alone but how the fitted Hamiltonian continues in the control parameter. To our knowledge, this is the first demonstration that a learned Hamiltonian can qualitatively extrapolate from predominantly regular dynamics into a broad chaotic sea absent from training. --- ### RAID: Towards Robust AI-Generated Image Detection with Bit-Reversed Images Authors: Renxi Cheng, Jie Gui, Hongsong Wang Published: 2026-07-31 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.28974v1 The rapid advancement of image generation models has made it increasingly difficult for people to distinguish AI-generated images from real ones. To prevent the potential risks associated with the misuse of fake images, AI-generated image detection has gained significant attention. Existing methods neglect the inherent differences between real and fake images, thus lacking robustness and generalization ability. In this work, we innovatively investigate AI-generated image detection using bit-planes, and introduce the bit-reversed image. We propose a simple yet effective pipeline consisting of construction of bit-reversed images, gradient-based patch selection and a convolutional classifier. Besides, we provide a theoretical analysis from the mathematical perspective to demonstrate the validity of our approach. We also introduce two challenging datasets for AI-generated image detection. Extensive experiments verify the effectiveness of our approach across different settings, including cross-generator generalization, cross-dataset generalization and zero-shot performance. Without bells and whistles, our approach outperforms existing methods on over 40 benchmarks, and is nearly 100 times faster than counterparts. The code is at https://github.com/renxi-seu/RAID. --- ### Don't Contrast the Impossible: Region-Constrained Batching for Contrastive User Modeling on a Local Community Platform Authors: Seungho Han, Byeongchang Kim, Jin Yu Published: 2026-07-31 Categories: cs.IR, cs.LG arXiv: https://arxiv.org/abs/2607.28971v1 Contrastive learning is widely used for user modeling in large-scale recommender systems, where standard in-batch negatives implicitly assume universal exposure that any user can be shown any item. On local community platforms such as Karrot, however, exposure is geographically constrained; many user-item pairs are impossible by design yet still treated as negatives during training, diluting the contrastive learning signal. We address this impossible negatives problem and propose Region-Constrained Batch Sampling (RCBS), a simple yet effective batching method that constructs region-homogeneous mini-batches so that users are contrasted primarily against items they could feasibly see. By replacing impossible negatives with feasible ones, RCBS naturally introduces harder and more informative negatives under realistic exposure constraints. With offline evaluations and online A/B tests, we show that RCBS consistently improves user representation quality and consequently enhances home feed ranking, retrieval, and display ads ranking. The resulting user embeddings have been deployed in production across various applications. --- ### LegoQ: Density-Matrix Representation Learning with Spectral-Spatial State Transitions for Hyperspectral Classification Authors: Weijia Cao, Xiaofei Yang, Fu Wang, Yicong Zhou, Xiang Zhou Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28970v1 Hyperspectral image classification is complicated by mixed pixels, spectral ambiguity, class imbalance, and limited annotations. Most current classifiers encode a pixel or patch as a deterministic vector and apply a linear or multilayer softmax head. Although effective for discrimination, this representation does not directly expose how mixed or uncertain a sample is. This paper presents \method, a classical density-matrix representation learning framework for hyperspectral images. The spectral bands are divided into groups and each group is mapped to a positive semi-definite, Hermitian, trace-normalized matrix state. A composable stack of spectral, spatial, and inter-group transitions then updates the states while repeatedly projecting them back to the valid state set. Instead of flattening the final features, \method\ aggregates the group states and compares them with learnable class-prototype density matrices through Uhlmann fidelity. The normalized eigenspectrum, von Neumann entropy, purity, and prototype fidelity provide sample-level diagnostics that are unavailable from a conventional vector head. On Indian Pines, ten runs yield an overall accuracy of $96.20\pm0.70\%$, an average accuracy of $95.57\pm1.29\%$, and a kappa coefficient of $95.66\pm0.80\%$. On WHU-Hi-LongKou, the best of ten runs reaches $97.52\%$ overall accuracy. Classification maps and feature projections show that the transition stack produces compact and better separated class structures. The results support constrained matrix-state learning as a practical alternative to vector-only hyperspectral classification without requiring quantum hardware. --- ### SafeNexus: Discovering and Steering Modality-Universal Safety Neurons in MLLMs Authors: Jian Yu, Fei Shen, Cong Wang, Jian Wang, Lu Jin, Xiaoyu Du, Jinhui Tang, Tat-Seng Chua Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28969v2 Although Large Language Models (LLMs) have demonstrated promising safety performance, extending them to Multimodal Large Language Models (MLLMs) exposes a significant gap between expanded multimodal capabilities and existing safety mechanisms. Current defenses remain predominantly confined to specific modal settings, thereby limiting their robustness against broader cross-modal threats. To bridge this gap, we introduce SafeNexus, a cross-modal safety alignment framework that adopts a dedicated neuron-level intervention strategy. First, we formulate a neuron localization paradigm that identifies functionally specialized neurons by characterizing intermediate-layer activation patterns and quantifying their functional salience through importance scoring. Building upon this paradigm, we exploit contrastive data to identify modality-bound safety neurons (BS-Neurons), and validate their role in regulating safety behavior within each modality via targeted suppression. Further cross-modal analysis defines modality-universal safety neurons (US-Neurons) as the shared subset of BS-Neurons identified across individual modalities, serving as the core for defending against harmful cross-modal attacks. We observe that suppressing these neurons substantially degrades safety performance across modalities, while leaving overall utility largely unaffected. Building on these insights, we propose two safety alignment strategies: activation-level safety amplifier and safety neuron calibrator. The proposed strategies enhance model safety through two distinct routes: the former amplifies the activation magnitudes of US-Neurons, while the latter selectively calibrates them via targeted fine-tuning. Extensive experiments demonstrate that our method outperforms prevailing state-of-the-art approaches on safety benchmarks spanning diverse modality combinations, while effectively preserving utility. --- ### SafeNexus: Discovering and Steering Modality-Universal Safety Neurons in MLLMs Authors: Jian Yu, Fei Shen, Cong Wang, Jian Wang, Lu Jin. Xiaoyu Du, Jinhui Tang, Tat-Seng Chua Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28969v1 Although Large Language Models (LLMs) have demonstrated promising safety performance, extending them to Multimodal Large Language Models (MLLMs) exposes a significant gap between expanded multimodal capabilities and existing safety mechanisms. Current defenses remain predominantly confined to specific modal settings, thereby limiting their robustness against broader cross-modal threats. To bridge this gap, we introduce SafeNexus, a cross-modal safety alignment framework that adopts a dedicated neuron-level intervention strategy. First, we formulate a neuron localization paradigm that identifies functionally specialized neurons by characterizing intermediate-layer activation patterns and quantifying their functional salience through importance scoring. Building upon this paradigm, we exploit contrastive data to identify modality-bound safety neurons (BS-Neurons), and validate their role in regulating safety behavior within each modality via targeted suppression. Further cross-modal analysis defines modality-universal safety neurons (US-Neurons) as the shared subset of BS-Neurons identified across individual modalities, serving as the core for defending against harmful cross-modal attacks. We observe that suppressing these neurons substantially degrades safety performance across modalities, while leaving overall utility largely unaffected. Building on these insights, we propose two safety alignment strategies: activation-level safety amplifier and safety neuron calibrator. The proposed strategies enhance model safety through two distinct routes: the former amplifies the activation magnitudes of US-Neurons, while the latter selectively calibrates them via targeted fine-tuning. Extensive experiments demonstrate that our method outperforms prevailing state-of-the-art approaches on safety benchmarks spanning diverse modality combinations, while effectively preserving utility. --- ### A robust association between LLM use and scientific productivity: Assessing stopping-time selection Authors: Keigo Kusumegi, Xinyu Yang, Paul Ginsparg, Mathijs de Vaan, Toby Stuart, Yian Yin Published: 2026-07-31 Categories: cs.DL, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2607.28968v1 Renault, Bergeaud, and Bosquet (hereafter RBB) argue that dating LLM adoption as the first month in which an author's abstract is flagged induces a stopping-time selection that can produce a positive event-study path even when there is no causal effect. Although this mechanism is mathematically possible, it does not constitute proof of a null effect. Recalibrating RBB's own random placebo to the detector's realized flag rate, we show that the measured association stays well above this benchmark, so the artifact is too small to explain the productivity changes. We further re-estimate the association between LLM adoption and productivity with a series of complementary designs in which the timing artifact cannot bias the estimate: a before-and-after comparison that dates adoption in one year and measures output in another, a conservative control group for difference-in-differences, an intensity-based specification that never defines an adoption date, and a rank-based measurement holding the flag rate fixed. A positive productivity association persists across all of these estimates, while the same tests run on pre-ChatGPT placebo data return null effects. The artifact RBB identify is real but bounded, and it does not account for the pattern we report. --- ### Visual Distribution Anchoring for Efficient Prompt Tuning Authors: Pouya Parsa, Raoof Zare Moayedi, Seongjin Choi Published: 2026-07-31 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2607.28967v1 Prompt tuning adapts vision--language models with few trainable parameters, but existing approaches trade off efficiency and adaptation: static textual prompts can overfit source classes, image-conditioned prompts add per-instance computation, and multimodal tuning modifies the visual branch. We propose VDA (Visual Distribution Anchoring), a training-free target adaptation framework that augments a frozen semantic classifier with class-level visual prototypes estimated offline from an unlabeled target pool. We first ask whether prototypes can be synthesized from class names. A text-to-centroid mapper reconstructs held-out source prototypes but fails under dataset shift because class names specify semantic identity, not target-domain appearance. An oracle analysis confirms that true target prototypes are highly discriminative. VDA therefore uses frozen semantic and domain-template classifiers to partition unlabeled target images into class-correlated groups. Confidence-ranked image features form normalized prototypes, fused with the semantic classifier using one global weight. Adaptation requires no target labels, target-side optimization, uniform class-prior assumption, iterative refinement, or test-query access, and yields a fixed, cacheable classifier. Controlled experiments show that class-specific partitioning drives gains and that visually local pseudo-label errors can remain useful despite being class-incorrect. Across ten ImageNet-to-target transfers, the same frozen design improves zero-shot CLIP, TCP, and MaPLe by 3.22, 3.39, and 3.35 points, respectively, improving nine of ten targets in every setting. Its visual correction further improves leakage-free PromptKD by 2.79 points, complementing zero-shot, source-prompted, multimodal-prompted, and target-distilled classifiers. --- ### BLADE: Boundary-Expanded and Layer-Adaptive Dynamic Exit for Efficient LLM Reasoning Authors: Keshu Fu, Keqin Peng, Jun Bai, Shuhan Qin, Chen Li, Junzhu Liang, Yefei Chen, Jiaqi Li, Yuanxin Ouyang Published: 2026-07-31 Categories: cs.CL arXiv: https://arxiv.org/abs/2607.28966v1 Large language models often improve task performance by generating long reasoning traces, but the resulting computation is frequently wasted on redundant verification and revision. Existing probe-based early-exit approaches mainly inspect explicit self-doubt expressions, leaving many earlier termination opportunities undetected. Expanding inspection to ordinary reasoning boundaries improves coverage, but also exposes highly diverse intermediate states whose predictive information may reside in different hidden layers. We present Boundary-Expanded and Layer-Adaptive Dynamic Exit for Efficient LLM Reasoning (BLADE), a lightweight framework that dynamically terminates reasoning by estimating whether the generated prefix is sufficient for correct answering. BLADE constructs multi-granular checkpoints from sentence, self-doubt, and paragraph boundaries, and derives robust training labels through repeated answer completions. It further learns a compact subset of informative probe layers instead of relying on fixed choices or expensive representations from all layers. At inference time, calibrated predictions are combined with checkpoint-specific confirmation rules to balance responsiveness and premature-exit risk. Experiments on five benchmarks and two Qwen3 reasoning models show that BLADE preserves near-baseline accuracy while reducing generated tokens by 24.8% on Qwen3-8B and 15.8% on Qwen3-4B. Ablation studies further confirm the benefits of diverse checkpoints and automatic layer selection, demonstrating an effective approach to more efficient LLM reasoning. --- ### Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates Authors: Weiyi He, Yuping Lin, Jiliang Tang, Yue Xing Published: 2026-07-31 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.28959v2 Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs). While existing mitigation strategies, e.g., latent adversarial training (LAT), have been developed, they still incur a high computational cost. In this work, focusing on LLM-based classifiers, we investigate computation-efficient strategies to speed up LAT from two complementary perspectives: (1) Defense-side optimization: We explore representation fine-tuning (ReFT) within LAT, and reveal a potential issue if there is a mismatch on which tokens to apply ReFT and the attack. (2) Attack-side optimization: When computing adversarial attacks in each LAT iteration, we extract only the relevant circuits from the LLM to construct a lightweight surrogate model, avoiding the computation in the forward-backward passes through the full model during the attack generation. For both perspectives, we provide theoretical justifications and numerical evidence to illustrate the effectiveness of the proposed strategies. Compared to standard LAT with full fine-tuning, our method on average reduces per-step adversarial-training FLOPs by 29.6% while requiring only 0.0118% trainable parameters, at a moderate cost in robustness. --- ### Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates Authors: Weiyi He, Yuping Lin, Jiliang Tang, Yue Xing Published: 2026-07-31 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.28959v1 Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs). While existing mitigation strategies, e.g., latent adversarial training (LAT), have been developed, they still incur a high computational cost. In this work, we comprehensively investigate computation-efficient strategies to speed up LAT from two complementary perspectives: (1) Defense-side optimization: We explore the representation fine-tuning (ReFT) within LAT, and reveal a potential issue if there is a mismatch on which tokens to apply ReFT and the attack. (2) Attack-side optimization: When computing adversarial attacks in each LAT iteration, we extract only the relevant circuits from the LLM to construct a lightweight surrogate model, avoiding the computation in the forward-backward passes through the full model during the attack generation. For both perspectives, we provide theoretical justifications and numerical evidence to illustrate the effectiveness of the proposed strategies. Ultimately, compared to standard LAT with full fine-tuning, our method on average reduces per-step adversarial-training FLOPs by 48.1% while requiring only 0.0118% trainable parameters. --- ### MerchantBench: Benchmarking LLM Agents for Long-Term Coherence in E-Commerce Operations Authors: Qiming Shi, Yulong Tao, Linbo Jin, Zhaolu Kang, Yibo Dou, Jiawen Zhu, Tianjun Pan, Shaokang Fu, Chengyu Wang, Siyue Li, Yaping Cheng, Di Weng et al. Published: 2026-07-31 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.28956v2 Large language model agents are increasingly evaluated as autonomous tool users, yet most benchmarks focus on bounded tasks with immediate success criteria. Real-world deployments often require Long-Term Coherence, the capacity to preserve purposeful behavior across extended horizons while adapting decisions to accumulated evidence. Evaluating this capacity requires a persistent environment in which actions constrain future choices, feedback arrives at heterogeneous delays, and incoherent behavior produces measurable cumulative effects. Seller-side e-commerce provides a suitable setting for this evaluation through recurrent and interdependent decisions over Product Sourcing, Listing and Pricing Control, Cash-Flow Management, and Mixed-Latency Feedback Adaptation. We introduce MerchantBench, a 365-day order-level simulation grounded in 98,843 real e-commerce product records and equipped with 26 tools for agent interaction. MerchantBench couples promptly observable Upstream Supplier Events with delayed Downstream Order Outcomes, requiring agents to follow individual order lifecycles and revisit earlier decisions. We evaluate eight LLMs under two agent frameworks in 48 runs, each spanning 365 simulated days. Our results reveal a substantial gap between even the latest LLMs and human participants, with the best LLM configuration attaining only 27.3\% of the mean final net assets achieved by human participants. --- ### Retrieval-Driven Training-Free AI-Generated Video Attribution Authors: Renxi Cheng, Chaolei Han, Jie Gui, Hongsong Wang Published: 2026-07-31 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.28955v1 AI-generated videos are becoming increasingly realistic and difficult to distinguish from authentic ones, which facilitates malicious misuse and poses growing threats to cybersecurity and social governance. Attributing AI-generated videos to their specific generative sources is therefore of critical importance for forensic investigation and legal regulation. However, most existing visual attribution methods focus on images and particularly rely on the image generation model, thereby lacking the ability to generalize to large-scale AI-generated video data. To address these limitations, we introduce an training-free AI-generated video attribution paradigm. Specifically, we formulates AI-generated video attribution as an instance retrieval task, and design a generative fingerprint-based pipeline. This pipeline consists of an adapted orthogonal color transformation, multi-scale quantized residual generation, and temporal-semantic aggregation, progressively capturing and integrating artifacts introduced by generative models across video frames. Extensive experiments on the GenVidBench benchmark demonstrate that our method achieves strong performance in both AI-generated video detection and attribution, outperforming existing state-of-the-art methods with a Rank-1 accuracy of 20.5% and a mean Average Precision of 16.6%. The code is at https://github.com/renxi-seu/Video_Attribution. --- ### Mining Verdict Boundaries for Neural Network Verification Authors: Jiawei Ren, Guanqin Zhang, Zhenya Zhang, Yulei Sui Published: 2026-07-31 Categories: cs.LG, cs.LO, cs.SE arXiv: https://arxiv.org/abs/2607.28954v1 Branch and Bound (BaB) aims to achieve complete verification of neural networks by adaptively partitioning the problem and applying off-the-shelf verifiers to subproblems. Its problem-splitting history can be represented as a tree, where each subproblem corresponds to a child node. A key problem of BaB lies in searching for the verdict boundaries across all the paths that divide the verified and unverified subproblems. We observe that the existing BaB approach tackles this problem by solving each expensive subproblem sequentially along the tree path as its depth increases, requiring costly bounds propagation at every visited BaB tree node (i.e., subproblem), which is inefficient. To address this issue, we propose effective search approaches that leverage the monotonicity of each path to efficiently and precisely locate the verdict boundary by simultaneously splitting multiple activation functions (e.g., ReLU), rather than processing them one at a time as in the classical approach. Our approach performs an effective exponential search along each path, allowing us to skip many boundary-unrelated subproblems when identifying the verdict boundary. The enhanced version further improves this process by estimating the boundary's position using quantitative information obtained from subproblem solving. We perform experimental evaluation on commonly-used benchmarks to assess our proposed techniques, and compare them with recent BaB-based approaches. --- ### Automated classification method of COVID-19 cases from chest CT volumes using 2D and 3D hybrid CNN for anisotropic volumes Authors: Masahiro Oda, Tong Zheng, Yuichiro Hayashi, Yoshito Otake, Masahiro Hashimoto, Toshiaki Akashi, Shigeki Aoki, Kensaku Mori Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28950v1 This paper proposes an automated classification method of chest CT volumes based on likelihood of COVID-19 cases. Novel coronavirus disease 2019 (COVID-19) spreads over the world, causing a large number of infected patients and deaths. Sudden increase in the number of COVID-19 patients causes a manpower shortage in medical institutions. Computer-aided diagnosis (CAD) system provides quick and quantitative diagnosis results. CAD system for COVID-19 enables efficient diagnosis workflow and contributes to reduce such manpower shortage. This paper proposes an automated classification method of chest CT volumes for COVID-19 diagnosis assistance. We propose a COVID-19 classification convolutional neural network (CNN) that has a 2D/3D hybrid feature extraction flows. The 2D/3D hybrid feature extraction flows are designed to effectively extract image features from anisotropic volumes such as chest CT volumes for diagnosis. The flows extract image features on three mutually perpendicular planes in CT volumes and then combine the features to perform classification. Classification accuracy of the proposed method was evaluated using a dataset that contains 1288 CT volumes. An averaged classification accuracy was 83.3%. The accuracy was higher than that of a classification CNN which does not have 2D and 3D hybrid feature extraction flows. --- ### Overcoming the Weakest-Link Effect in LLM-Driven Program Optimization via Heterogeneous Edit Recombination Authors: Jingwen Fu, Zhen Liu, Yuhan Liu, He Zhang, Nanning Zheng Published: 2026-07-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.28947v1 Large language models (LLMs) are increasingly used to solve complex problems by searching over program space, offering a general paradigm for scientific problems that can be naturally represented and solved as programs. Despite recent progress, identifying effective optimization directions for a candidate program remains challenging. By analogy with automatic differentiation, existing methods typically guide the search using a textual ``gradient'': a first-order update direction expressed as textual edits. Such gradients are inferred either from previously evaluated programs or from LLM-generated feedback on the implicit program-score mapping. However, these estimates become increasingly unreliable as the program--score mapping grows more complex, limiting their practical utility. We argue that explicit gradients are not essential for effective program optimization. Leveraging their prior knowledge, LLMs can propose plausible atomic edits directly from the current program, thereby enabling a zeroth-order optimization strategy. However, zeroth-order search suffers from a \textit{weakest-link effect}: when a bundle of edits is accepted or rejected as a whole, a single harmful edit can negate the benefits of all remaining edits. To address this issue, we introduce HERO, a program optimizer that prompts an LLM to generate diverse, non-overlapping atomic edits and then systematically selects and composes them into coherent program improvements using evaluator scores. We evaluate HERO across algorithmic problems, strategy games, the design of LLM-based agentic systems, and robotic path planning. Across these domains, HERO consistently discovers higher-scoring programs and converges substantially faster than prior LLM-based optimizers, while consuming fewer tokens. --- ### Shapley-Value-Based Feature Attribution for Data Masking Authors: Xinxue, Qu, Francis Bilson Darku, Hong Guo Published: 2026-07-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.28946v1 Despite its many benefits, widespread access to individuals' personal data also causes severe privacy concerns for consumers, companies, and policymakers. This study proposes a novel framework that adapts the Shapley-value-based feature attribution approach to the problem domain of data privacy by capturing the two crucial dimensions of data privacy---disclosure risk and data utility. Our proposed framework takes a holistic view of data masking through a fair feature attribution approach based on Shapley values. Different from the existing literature that mostly focuses on the risk-utility tradeoff at the dataset level, the proposed framework addresses the tradeoff at the feature level. Furthermore, the proposed framework is agnostic to data masking methods, statistical and machine learning methods, and data utility and disclosure risk evaluation metrics. Experimental results show that our proposed method can effectively reduce disclosure risk while preserving data utility. --- ### FairDiffuseVQVAE: Sampling-Time Fairness in Tabular Diffusion via Conditional Refinement of Vector-Quantized Latents Authors: Nitish Nagesh, Mahdi Bagheri, Amir M. Rahmani Published: 2026-07-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.28945v1 Synthetic tabular data is increasingly used in privacy-preserving data sharing, data augmentation, and to mitigate downstream classifier bias. State-of-the-art tabular diffusion models such as TabDDPM and TabSyn achieve excellent distributional fidelity but offer no mechanism for fairness; conversely, fairness-aware tabular generators (DECAF, FairTGAN, FairTabDDPM) impose explicit fairness penalties at training time, yielding modest fairness gains at substantial cost to either sample quality or downstream utility. We introduce FairDiffuseVQVAE, a two-stage architecture that decouples fidelity from fairness: a vector-quantized autoencoder with a row-level discriminator (Stage~1, no fairness terms) is followed by a DiffuseVAE-style continuous diffusion refiner that conditions on both the Stage-1 reconstruction and the protected attribute via classifier-free guidance (Stage~2). Fairness emerges as a property of the sampling distribution -- uniform sampling of the protected attribute at inference time enforces demographic parity by construction, rather than from competing loss terms. On the Adult, Bank and COMPAS datasets, FairDiffuseVQVAE achieves the highest mean Demographic Parity Ratio ($0.702$, $+47\%$ over FairTabDDPM) and Equalized Odds Ratio ($0.686$, $+100\%$). It also attains the lowest mean pair-wise correlation error ($0.034$) of any published method, while explicitly trading $\sim$$15$ AUC points for these fairness gains. --- ### A Biometric Sensor Network to Enable Real-Time Measurement of Individual Student Engagement in STEM Lecture Environments Authors: Ahmed Elsayed Published: 2026-07-31 Categories: cs.CR, cs.CV, cs.SE, eess.IV arXiv: https://arxiv.org/abs/2607.28944v1 Student engagement (SE) is a critical predictor of academic performance and retention in STEM education, yet existing measurement approaches are often intrusive, manually intensive, or unsuitable for real-time classroom use. This thesis proposes a novel $\textit{Biometric Sensor Network}$ (BSN) designed to enable real-time measurement and continuous tracking of individual student engagement in STEM classroom environments. The system enables capturing of behavioral, emotional, and cognitive indicators through camera-based sensing while preserving ethical and privacy constraints. To measure these indicators unobtrusively and ethically, we propose a BSN composed of $\textit{Student Processing Units}$ (SPUs) that function as distributed sensing nodes. The network is explicitly designed to satisfy five objectives: it must be $\textbf{non-intrusive}, \textbf{non-invasive}, \textbf{non-stigmatizing}, \textbf{real-time}$, and $\textbf{automatic}$, while ensuring rigorous protection of student data security and privacy. Each SPU supports two operational modes: (i) a $\textit{dataset-collection mode}$, in which raw student video is temporarily recorded to construct a private SE dataset for model training and validation, and (ii) an $\textit{analysis mode}$, in which the SPU performs real-time inference on 10-second video segments without storing or transmitting raw frames. In this analysis role, each SPU enables fully on-device processing---including face detection, gaze estimation, and affective analysis---ensuring that no identifiable video data leaves the device. A secure backend infrastructure manages device authentication, session orchestration, and encrypted data ingestion. The full system integrates hardware design, computer-vision pipelines, wireless networking, security protocols, and session-level data management. --- ### NeSyFS: A Neuro-symbolic Fast-Slow Thinking Framework for LLM Agent under Partial Observability Authors: Duo Xu, Faramarz Fekri Published: 2026-07-31 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.28942v3 Recently Large Language Models (LLMs) have been increasingly deployed as autonomous agents in applications such as self-reflection, retrieval-augmented generation, and scientific discovery. In these settings, agents must act based on limited observations rather than full environmental states, leading to partial observability. This introduces several key challenges: belief state inference, task objective misalignment, and planning under uncertainty. Prior approaches typically condition actions on full or summarized action-observation histories whose redundant and irrelevant information can mislead the decision making of LLM agent. Inspired by human cognition, we propose a novel neuro-symbolic fast-slow thinking (NeSyFS) framework for LLM agent, addressing the challenges introduced by partial observability in a unified approach. We use a knowledge graph (KG) to represent the belief state, providing triplets as context for every module of NeSyFS. The fast-thinking module performs reactive action, while slow-thinking conducts a new uncertainty-aware planning by following the high-level structure of twisted sequential Monte Carlo (TSMC) algorithm. To mitigate the misalignment of task objective, a reflection module is used to reflect fast-thinking actions, and also switches to the slow-thinking module whenever reactive actions repeatedly fail. Experiments on three representative benchmarks, i.e. ALFWorld, Webshop, and ScienceWorld, demonstrate significant advantages over previous methods. --- ### NeSyFS: A Neuro-symbolic Fast-Slow Thinking Framework for LLM Agent under Partial Observability Authors: Duo Xu, Faramarz Fekri Published: 2026-07-31 Categories: cs.AI arXiv: https://arxiv.org/abs/2607.28942v2 Recently Large Language Models (LLMs) have been increasingly deployed as autonomous agents in applications such as self-reflection, retrieval-augmented generation, and scientific discovery. In these settings, agents must act based on limited observations rather than full environmental states, leading to partial observability. This introduces several key challenges: belief state inference, task objective misalignment, and planning under uncertainty. Prior approaches typically condition actions on full or summarized action-observation histories whose redundant and irrelevant information can mislead the decision making of LLM agent. Inspired by human cognition, we propose a novel neuro-symbolic fast-slow thinking (NeSyFS) framework for LLM agent, addressing the challenges introduced by partial observability in a unified approach. We use a knowledge graph (KG) to represent the belief state, providing triplets as context for every module of NeSyFS. The fast-thinking module performs reactive action, while slow-thinking conducts a new uncertainty-aware planning by following the high-level structure of twisted sequential Monte Carlo (TSMC) algorithm. To mitigate the misalignment of task objective, a reflection module is used to reflect fast-thinking actions, and also switches to the slow-thinking module whenever reactive actions repeatedly fail. Experiments on three representative benchmarks, i.e. ALFWorld, Webshop, and ScienceWorld, demonstrate significant advantages over previous methods. --- ### Latent Lie-Poisson Neural Networks (LLPNNs): Discovering the motion of Lie-Poisson systems through observable data and latent dynamics Authors: Vakhtang Putkaradze Published: 2026-07-31 Categories: cs.LG, math.DS, math.OC arXiv: https://arxiv.org/abs/2607.28939v1 Structure-preserving neural networks are essential for the long-term prediction of Hamiltonian systems from data. Many important Hamiltonian systems in mechanics and control admit symmetry reduction to Lie--Poisson systems, including rigid bodies, underwater vehicles, fluids, plasmas, and optimal control problems. A fundamental challenge in learning such systems is that their dynamics evolve in momentum variables that are typically unobservable, while available data consist only of observable quantities such as configurations and velocities. In optimal control applications, the situation is further complicated because the latent variables contain unobservable co-states and the Hamiltonian may be degenerate, preventing the existence of a corresponding Lagrangian and rendering the encoder-decoder approaches inapplicable. We introduce Latent Lie--Poisson Neural Networks (LLPNNs), a structure-preserving framework for learning Lie--Poisson dynamics directly from observable data. The proposed approach exploits three geometric ingredients: (i) learning either a Hamiltonian decoder or a pseudo-Lagrangian encoder on the active variables, (ii) constructing latent trajectories through a universal Noether invariant arising from Lie--Poisson symmetry reduction, and (iii) reconstructing observable and latent dynamics through Lie--Poisson flows combined with Magnus-based Lie-group updates. The resulting method preserves the geometric structure and is applicable to both regular and degenerate Hamiltonian systems. We demonstrate the method on three examples: a generalized rigid body on SO(3), Kirchhoff's underwater vehicle on SE(3), and an optimal-control problem for interacting vehicles on $SE(2)^N$. Numerical experiments show excellent long-term predictive accuracy, strong robustness to noise, and competitive performance using only modest datasets and lightweight neural-network architectures. --- ### DiffAttack: Evasion Attacks Against Face Recognition via Latent Diffusion Models Authors: Omid Ahmadieh, Nima Karimian Published: 2026-07-31 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2607.28936v1 Facial biometric identification relies on the distinctiveness of user attributes within a high-dimensional embedding space. However, the decision boundaries of deep face recognition (FR) systems are often sufficiently narrow that they can be conflated, rendering the models vulnerable to adversarial attacks. In such scenarios, the FR system fails to distinguish between an authentic source and a meticulously crafted adversarial face. Existing adversarial methods targeting facial biometrics are limited in both performance and their ability to generate high-quality images that are imperceptible to humans. Moreover, these methods often fail when the source and target images belong to different demographic groups or genders. To address these limitations, we present a novel approach for adversarial face generation via latent-space optimization. We leverage latent diffusion models directly to guide generation toward target identity embeddings, as measured by a face recognition model. Our proposed \textbf{DiffAttack} framework has been evaluated on standard benchmarks, such as the FFHQ and CelebA-HQ datasets. DiffAttack significantly outperforms existing adversarial techniques, achieving a high average attack success rate of 84.86% across multiple face recognition models (e.g., FaceNet). Notably, DiffAttack demonstrates superior transferability, surpassing traditional noise-based methods by over 15.28% and semantic-based approaches by approximately 5.21% on benchmark datasets like FFHQ and CelebA-HQ. --- ### Group-wise Supervision with Focal-Dice Loss for Long-Tailed Indoor Semantic Occupancy Prediction Authors: Qi Zheng, Zihuang Su, Xiao Pan Published: 2026-07-31 Categories: cs.CV arXiv: https://arxiv.org/abs/2607.28935v1 Recently, 3D semantic occupancy prediction has garnered increasing attention for understanding the indoor scene. However, unlike structured outdoor environments, indoor scenes feature a high diversity of object categories that exhibit a severe long-tailed distribution, which has become a core bottleneck limiting the performance of existing models. To tackle this challenge, we propose a novel method, Group-UFD Occ, based on hierarchical semantic supervision and synergistic loss optimization. At the architectural level, we introduce a fine-grained semantic grouping strategy and design multi-scale, parallel ``main-expert'' prediction heads to guide the model in efficiently learning tail-class features through deep regularization. At the optimization level, we introduce the Unified Focal-Dice (UFD) loss. This synergistic loss function dynamically focuses on hard samples at the per-voxel level. Meanwhile, it simultaneously optimizes the geometric integrity of predicted objects from a region-based perspective. We conducted experiments on the large-scale EmbodiedScan dataset. The results demonstrate that our method yields a relative improvement of 11.38\% over the baseline, with substantial accuracy gains in several critical long-tailed categories. --- ### FairFund-Bench: Evaluating Distributive Bias in LLM Resource Allocation Authors: Martin Lukk Published: 2026-07-31 Categories: cs.CL, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2607.28934v2 Large language models (LLMs) are increasingly involved in the distribution of scarce resources, raising concerns about biased allocations based on characteristics like race and gender. Recent LLM audits have produced inconsistent results, however, finding evidence of both positive and negative discrimination towards women and ethnic minorities, even for the same models. We show that this disagreement can arise from differences in audit format and introduce FairFund-Bench, a benchmark that systematically varies key features of previous audit designs: the evaluation task (rating, ranking, or allocation), comparison context (single or multi-stimulus), and whether the audit is transparent or disguised. The benchmark comprises 600 English-language requests for financial assistance created from human-authored templates (calibrated against 1.3M real GoFundMe campaigns) across three domains, four race and two gender categories, and five causal framings of need derived from welfare deservingness theory. Across 14 models, audit format changes the direction of bias: models advantage minorities when rating claimants individually but penalize some groups when ranking them side by side. Bias magnitude, though small overall, is several times greater in disguised audits than in transparent ones, where, faced with appeals differing only in claimants' names, models overwhelmingly split funds equally. Causal framing effects, by contrast, exceed demographic effects by roughly an order of magnitude and are consistent across models and audit formats, indicating that current LLMs robustly reproduce human deservingness evaluations. The benchmark scores models on four criteria (demographic bias, deservingness alignment, cross-task consistency, and cross-context consistency), is publicly available, and can be readily adapted to other substantive domains. --- ### FairFund-Bench: Evaluating Distributive Bias in LLM Resource Allocation Authors: Martin Lukk Published: 2026-07-31 Categories: cs.CL, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2607.28934v1 Large language models (LLMs) are increasingly involved in the distribution of scarce resources, raising concerns about biased allocations based on characteristics like race and gender. Recent LLM audits have produced inconsistent results, however, finding evidence of both positive and negative discrimination towards women and ethnic minorities, even for the same models. We show that this disagreement can arise from differences in audit format and introduce FairFund-Bench, a benchmark that systematically varies key features of previous audit designs: the evaluation task (rating, ranking, or allocation), comparison context (single or multi-stimulus), and whether the audit is transparent or disguised. The benchmark comprises 600 requests for financial assistance created from human-authored templates (calibrated against 1.3M real GoFundMe campaigns) across three domains, four race and two gender categories, and five causal framings of need derived from welfare deservingness theory. Across 14 models, audit format changes the direction of bias: models advantage minorities when rating claimants individually but penalize some groups when ranking them side by side. Bias magnitude, though small overall, is several times greater in disguised audits than in transparent ones, where, faced with appeals differing only in claimants' names, models overwhelmingly split funds equally. Causal framing effects, by contrast, exceed demographic effects by roughly an order of magnitude and are consistent across models and audit formats, indicating that current LLMs robustly reproduce human deservingness evaluations. The benchmark scores models on four criteria (demographic bias, deservingness alignment, cross-task consistency, and cross-context consistency), is publicly available, and can be readily adapted to other substantive domains. --- ### Learning Optimal Dynamic Matching via Graph Neural Networks Authors: Genta Okada, Shunya Noda, Junpei Komiyama, Akira Matsushita Published: 2026-07-31 Categories: cs.LG, cs.GT, econ.TH arXiv: https://arxiv.org/abs/2607.28925v2 Dynamic matching markets require decisions about whom to match and when: matching now yields value but removes participants who may create better future opportunities. We develop a value-based reinforcement-learning framework for this problem on finite, evolving weighted graphs. We study an infinite-horizon continuous-time model with stochastic arrivals, node-type transitions, edge realizations, and exogenous exits. We prove an event-time reduction: without loss of optimality, the planner acts immediately after each exogenous event and then waits for the next one. We further show that the optimal edge-wise $Q$-function is characterized by a single continuation-value function on post-decision residual graphs, reducing the learned object from state-action values to graph values. Exact action selection still requires combinatorial matching optimization; we approximate the value with a graph neural network, train it by temporal-difference learning, and use it in a forward-greedy matching heuristic. In a binary-type benchmark, the learned policy substantially outperforms immediate and threshold-greedy rules by preserving common nodes for rare arrivals of valuable matches while forming lower-value matches only in thick pools. In a kidney paired donation benchmark, it performs similarly to immediate greedy when exits are unpredictable, recovers the logic of patient matching when warnings are reliable, and outperforms the better of Immediate Greedy and Patient Greedy across intermediate warning probabilities. These results show that residual-graph value learning yields state-dependent dynamic matching policies that adapt to realized connectivity and exit information. --- ### Learning Optimal Dynamic Matching via Graph Neural Networks Authors: Genta Okada, Shunya Noda, Junpei Komiyama, Akira Matsushita Published: 2026-07-31 Categories: cs.LG, cs.GT, econ.TH arXiv: https://arxiv.org/abs/2607.28925v1 Dynamic matching markets require decisions about whom to match and when: matching now yields value but removes participants who may create better future opportunities. We develop a value-based reinforcement-learning framework for this problem on finite, evolving weighted graphs. We study an infinite-horizon continuous-time model with stochastic arrivals, node-type transitions, edge realizations, and exogenous exits. We prove an event-time reduction: without loss of optimality, the planner acts immediately after each exogenous event and then waits for the next one. We further show that the optimal edge-wise $Q$-function is characterized by a single continuation-value function on post-decision residual graphs, reducing the learned object from state-action values to graph values. Exact action selection still requires combinatorial matching optimization; we approximate the value with a graph neural network, train it by temporal-difference learning, and use it in a forward-greedy matching heuristic. In a binary-type benchmark, the learned policy substantially outperforms immediate and threshold-greedy rules by preserving common nodes for rare arrivals of valuable matches while forming lower-value matches only in thick pools. In a kidney paired donation benchmark, it performs similarly to immediate greedy when exits are unpredictable, recovers the logic of patient matching when warnings are reliable, and outperforms the better of Immediate Greedy and Patient Greedy across intermediate warning probabilities. These results show that residual-graph value learning yields state-dependent dynamic matching policies that adapt to realized connectivity and exit information. --- ### Gated Q-learning: Add Off-Policy Bias to Taste Authors: Brett Daley Published: 2026-07-31 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2607.28916v1 Multistep credit assignment is critical for sample-efficient reinforcement learning, yet managing off-policy bias in Q-learning remains a fundamental challenge. For 30 years, practitioners have been limited to a binary choice: eliminate the bias at the cost of severely truncated eligibility traces (Watkins' Q($λ$)), or ignore the bias to learn faster while injecting detrimental errors into the value estimates (Peng's Q($λ$)). Modern off-policy estimators fail to resolve this tension, as importance-sampling ratios collapse under Q-learning's greedy target policy. We introduce Gated Q-learning, a novel algorithmic framework that ends this dilemma by smoothly interpolating between the two historical extremes. Rather than relying on importance sampling, our approach employs a continuous, state-action-dependent gating mechanism to selectively attenuate eligibility traces in an exploration-aware manner. We provide a rigorous theoretical foundation for this mechanism, proving that the expected operator remains a contraction mapping and deriving its exact fixed point. Empirical evaluations verify that intermediate gating safely enables longer credit-assignment horizons, yielding faster initial learning than either extreme. Gated Q-learning offers a simple alternative to importance sampling while enabling customization of the effective multistep horizon and the amount of off-policy bias in Q-learning agents. --- ### Reflection or Re-Generation? Why LLM Revision Fails Where Human Revision Succeeds Authors: Yefan Tao, Gerald Friedland, Madhusudhanan Chandrasekaran, Luyang Kong Published: 2026-07-31 Categories: cs.LG arXiv: https://arxiv.org/abs/2607.28908v1 Reflection, the ability to revisit and revise prior reasoning, is central to how humans improve their answers. Large language models (LLMs) are increasingly prompted to "reflect," yet whether this resembles human revision remains unclear. We introduce the Human-LLM Reflection Framework (HRF), a controlled two-pass protocol comparing human and LLM revision under identical conditions across self-, peer-, and cross-agent settings. Using an information-theoretic analysis based on per-iteration cross-entropy reduction, we find two failure modes of LLM reflection. On objective tasks with finite answer spaces, reflection yields near-zero information gain (Delta I approx 0), behaving as neutral re-generation indistinguishable from re-sampling. On subjective tasks, it yields significant negative gain (Delta I < 0), moving predictions away from the target. Human revision, by contrast, yields positive gain in both settings. Cross-agent experiments localize the failure to the revision step, not input quality: LLMs degrade even high-quality human responses. Diagnostic analyses (revision conditioned on first-pass correctness, and oracle-guided revision against a random-reshuffle baseline) show that which sub-step dominates varies by task and by model rather than reducing to a single mechanism: self-error detection is present on objective multiple-choice tasks but weak on subjective ones, and recovery under an oracle error signal exceeds the baseline for some models and falls below it for others. The unifying account is structural: without external information, self-conditioned revision cannot reduce uncertainty about the target, so LLM reflection is better understood as conditioned re-generation than as genuine error-driven revision. --- ### Domain-Adaptive Deep Joint Source-Channel Coding for Image Classification Authors: Yishen Li, Xuechen Chen, Xiaoheng Deng, Hao Zhang Published: 2026-07-31 Categories: cs.IT, cs.CV, eess.SP arXiv: https://arxiv.org/abs/2607.28907v1 Deep joint source--channel coding (Deep JSCC) enables visual semantic transmission by mapping inputs directly to channel symbols and task outputs, but its performance can deteriorate under distribution shifts between training and deployment domains. We study single-source domain adaptation for task-oriented Deep JSCC and formulate a classification-capacity-invariance (CCI) function to characterize how the available channel capacity and class-conditional cross-domain invariance affect target domain classification accuracy. A scalar linear analysis of source-domain-optimal solutions and a controlled shallow nonlinear validation show that target domain classification accuracy can vary non-monotonically with the invariance constraint and with available capacity along separate control paths obtained by varying the transmitted dimension or CSNR. We then propose a domain-adaptive Deep JSCC framework that combines pseudo-label-based class-level adversarial alignment with supervised contrastive learning on confidence-filtered target samples. Experiments on digit and PACS datasets over AWGN and Rayleigh fading channels demonstrate improved target domain generalization without introducing additional inference-time networks. On SVHN $\rightarrow$ MNIST, the proposed method achieves 98.15\% target-domain accuracy at a CSNR of 10 dB. --- ### Token-Level Diagnosis of Sycophancy in LLMs with Attribution-Guided Steering Authors: Hieu Nguyen, Mahammed Kamruzzaman, Anshuman Chhabra, Gene Louis Kim Published: 2026-07-31 Categories: cs.CL arXiv: https://arxiv.org/abs/2607.28906v1 Sycophancy refers to the tendency for large language models (LLMs) to match user beliefs at the cost of factual correctness, thereby undermining model reliability. Prior work on evaluating sycophancy in LLMs aims to assess whether a model's output matches an authority's claim, but cannot reveal which part of the prompt drives this sycophantic behavior. To bridge this gap, we investigate the relationship of sycophantic responses with an authority's credentials, their assertive claim, and the problem statement. We introduce the Authority Share Index (ASI), an Integrated Gradients-based token attribution method, which measures the degree to which a model's decision is driven by authority-related text. Through extensive experiments across five models and 30 test configurations, we find that sycophantic responses consistently direct more attention toward authority tokens than resistant ones. Moreover, our token attribution method reveals that for the sycophantic cases, the claim asserted by the authority receives more attention than the authority's credentials. Building on these findings, we propose attribution-guided contrastive activation steering to mitigate LLM sycophancy. Our method constructs a steering vector from high-attribution tokens of sycophantic and resistant responses, selectively pushing models toward resistance. This enables inference-time steering without retraining, lowering sycophancy from 96% to 25% in the strongest case. Together, our results show that token-level attribution can both explain what drives sycophancy and directly inform a practical intervention. --- ### Design Concept: Scaffolding Geopolitical Reflection Among Tech Workers Authors: Sydney Reis Published: 2026-07-31 Categories: cs.HC, cs.AI, cs.CY arXiv: https://arxiv.org/abs/2607.28904v1 This paper presents a speculative Human-Computer Interaction design proposal for encouraging geopolitical reflexivity amongst tech workers at geopolitically relevant technology companies. Recent scholarship in International Relations and Science and Technology Studies increasingly recognizes technology firms and their workers as geopolitical actors whose decisions shape international dynamics. However, existing Responsible Innovation and Responsible AI approaches rarely engage with the geopolitical narratives and imaginaries that underpin contemporary AI development. Building upon RI scholarship on reflexivity, reflective HCI, and creative HCI work on computational narratives, this paper proposes an AI-enabled interactive narrative system in which users engage with a speculative scenario centred on technology, power, and geopolitics. Through narrative interaction, archetype assignment, and socially scaffolded workshop reflection, the system aims to encourage target users to critically examine their assumptions, values, and positionality within broader sociotechnical systems. We argue that speculative narrative systems may offer a productive avenue for introducing geopolitical reflexivity into responsible technology initiatives without relying on prescriptive or moralising approaches. --- ### Structured Neural Chaos: An Adaptive Surrogate Modeling Framework for Functional Uncertainty Quantification and Global Sensitivity Analysis Authors: Isabel Corona Guevara, Yeping Hu Published: 2026-07-31 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2607.28903v1 Variance-based global sensitivity analysis (GSA) plays a key role in uncertainty quantification by identifying the contributions of uncertain inputs to the variability of the model response. The repeated model evaluations required for these tasks are often prohibitively expensive; surrogate models provide an efficient alternative by constructing inexpensive approximations of the underlying system response. Constructing surrogate models that combine scalability and interpretability for systems with high-dimensional stochastic inputs and functional responses remains challenging, particularly when sensitivity estimates are required across spatial or temporal domains. Polynomial chaos expansion (PCE) provides an effective framework for uncertainty propagation and sensitivity analysis due to its orthogonal structure and direct relationship with variance-based sensitivity measures. However, PCE suffers from the curse of dimensionality, whose computational burden is amplified for problems with functional responses. In this work, we introduce the Structured Neural Chaos (sNC) expansion as a surrogate modeling framework for variance-based GSA, inspired by the interpretability and orthogonal structure of PCE. The proposed framework retains the interpretability of structured decompositions while leveraging the expressive power of neural networks. The sNC expansion mirrors a truncated functional ANOVA decomposition, where each interaction component admits a separable low-rank approximation whose basis functions and coefficients are parameterized by neural networks. The expansion is constructed sequentially, adaptively identifying the dominant modes within each ANOVA subspace and determining the effective complexity of the representation. The resulting structure enables the extraction of statistical and sensitivity quantities directly from the coefficients of the sNC expansion at negligible cost. --- ### Adaptivity via a Parallel Architecture for Stochastic Gradient Methods Authors: Bin Fu Published: 2026-07-31 Categories: cs.LG, cs.AR, math.OC arXiv: https://arxiv.org/abs/2607.28902v2 Let $\mathrm{A}(x_0,y)$ be an algorithm with two inputs: an initial point $x_0$ and an integer parameter $y$, which specifies that $\mathrm{A}(.,.)$ executes at most $y$ iterations or steps. Given an integer $p\ge 1$, $p$ parallel processors search an appropriate value of $T$ for for $A(.)$. Each processor executes an infinite sequence of stages indexed by $i=0,1,2,\ldots$. At stage $i$, processor $j$ is assigned $T_{j,i}=h(j,i),$ where $h:\mathbb{N}\times\mathbb{N}\rightarrow\mathbb{R}^{+}$ is a prescribed function. Processor $j$ $(j=0,1,\ldots,p-1)$ then executes $\mathrm{A}(x_0,T_{j,i})$. The efficiency of the parallel framework is characterized by its $(p,α_p)$-approximation guarantee. Specifically, for every integer $T\ge T_0$, there exist a processor $j$ and a stage $i$ such that $T\le T_{j,i}\le T_{j,i}^*<α_p T,$ where $T_{j,i}^*=\sum_{t=0}^{i}T_{j,t}$ denotes the cumulative number of iterations executed by processor $j$ from the beginning to stage $i$. We prove that this framework achieves a $(p,α_p)$-approximation, and a tight lower bound for $α_p$ for all large $p$. We develop arithmetically simple stochastic gradient methods in which every division is of the form $x/2^t$ for some integer $t$, and integrate them into the proposed parallel framework. --- ### Adaptivity via a Parallel Architecture for Stochastic Gradient Methods Authors: Bin Fu Published: 2026-07-31 Categories: cs.LG, cs.AR, math.OC arXiv: https://arxiv.org/abs/2607.28902v1 We develop a parallel framework that assembles static gradient methods to achieve better adaptivity. A static gradient method, denoted by $\mathrm{GD}(x_0,T)$, takes as input an initial point $x_0\in\mathbb{R}^n$ and $T\in \mathbb{R}^+$ specifying the number $\floor{T}$ of iterations. The step size is chosen as $s=S(T)$, where $S(\cdot)$ is a predetermined function of $T$. The method then performs the iterations $ x_{i+1}=x_i-\fracη{s}\cdot g_i,$ where $g_i$ is a stochastic gradient evaluated at $x_i$, and $η$ is a scaling factor. For an integer $p\ge1$, the $p$ processors in the proposed parallel framework search for an appropriate value of $T$ according to a geometric sequence so that the resulting gradient descent satisfies the desired convergence conditions. Each processor executes an infinite sequence of stages indexed by $i=1,2,\ldots$. At stage $i$, processor $j$ is assigned $ T_{j,i}=h(j,i),$ where $h:\mathbb{N}\times\mathbb{N} \rightarrow\mathbb{R}^{+}$ is a prescribed function. Processor $j$ $(j=0,1,\ldots,p-1)$ executes $\mathrm{GD}(x_0, T_{j,i})$ at stage $i$. --- ### Curriculum-Based Noise Adaptation for Phoneme-to-Text Reconstruction in Visual Speech Recognition Authors: Matthew Kit Khinn Teng, Haibo Zhang, Takeshi Saitoh Published: 2026-07-30 Categories: cs.CL, cs.CV, cs.SD arXiv: https://arxiv.org/abs/2609.20839v1 Phoneme-centric visual speech recognition reconstructs sentences from intermediate phoneme predictions, making overall recognition performance highly dependent on the robustness of the phoneme-to-text reconstruction model. Existing reconstruction approaches are commonly trained on clean phoneme sequences or synthetically corrupted inputs, leading to a mismatch between training conditions and the realistic phoneme prediction errors encountered during inference. To address this limitation, this paper proposes progressive error curriculum training (PECT). This curriculum learning framework progressively adapts a No Language Left Behind (NLLB)-based phoneme-to-text reconstruction model using synthetic phoneme perturbations, multi-domain pseudo-labels, and target-domain pseudo-labels generated by a visual speech recognizer. By gradually exposing the reconstruction model to increasingly realistic phoneme prediction errors, the proposed framework improves robustness while preserving sentence-reconstruction accuracy. Experiments on the LRS2 and LRS3 benchmarks demonstrate that PECT consistently improves reconstruction performance across multiple phoneme-based visual speech recognition frontends, including visual automatic speech recognition (V-ASR), point visual automatic speech recognition (PV-ASR), and head-pose-aware visual speech recognition (HP-VSR) variants. In particular, PECT reduces the word error rate (WER) of HP-VSR-FiLMFuse (L4) from 23.3% to 22.2% on LRS2 and reduces the WER of HP-VSR-ResFiLM from 30.3% to 29.7% on LRS3. Comprehensive ablation studies and qualitative analyses further demonstrate the effectiveness of progressively adapting the reconstruction model to realistic phoneme prediction errors. These results show that PECT provides an effective and generalizable curriculum learning strategy for phoneme-to-text reconstruction in phoneme-centric visual speech recognition. --- ### Diagnose, Recover, Certify: Task Readiness under Hidden Dynamics Changes Authors: Nguyen Viet Tuan Kiet, Huynh Thi Thanh Binh Published: 2026-07-30 Categories: cs.AI arXiv: https://arxiv.org/abs/2609.20304v1 A deployed control policy can conceal consequential dynamics changes: an actuator may lose effectiveness without affecting the current task when the policy rarely excites it, despite being critical for a future task that has not yet been specified. We introduce task readiness under dormant dynamics drift, a decision problem that unifies active change diagnosis and post-change control recovery under a limited, task-agnostic interaction budget. An agent must identify whether and where local dynamics have changed, use a small number of informative interactions to characterize the change before downstream task identity is revealed, and subsequently provide each candidate task with either a recovered policy and a calibrated lower bound on its achievable return or an abstention decision to a safe fallback. We propose Evidence-Gated Matched-Pulse Transport, an intervention-based Bayesian procedure that couples fault localization with estimation of actuator effectiveness through a shared matched-response representation, thereby preserving diagnostic reliability while converting localized evidence into recovery-relevant uncertainty. This uncertainty is propagated to task-conditioned policy selection and readiness certification, enabling deployment decisions that explicitly trade off expected performance, confidence, and fallback use. We evaluate the resulting framework on a diverse suite of dormant-actuator benchmarks spanning multiple simulators, under a protocol that separates diagnosis from capability recovery, scores deployment by readiness coverage, selective risk, and interaction cost as well as return, and identifies the fault regimes in which transported evidence is decisive. --- ### Viveka-Insight: a cross-lingual concept graph and citation-grounded retrieval resource over the complete works of Swami Vivekananda in English and Bengali Authors: Tamal Maharaj Published: 2026-07-30 Categories: cs.CL, cs.DL, cs.IR arXiv: https://arxiv.org/abs/2609.20303v1 Classical philosophical corpora pose three compounding challenges for language resources: they exist in several languages without parallel alignment, their vocabulary is remote from that of contemporary readers, and generated text over culturally sensitive material must be verifiably grounded. We present Viveka-Insight, a bilingual resource and open-source pipeline for the works of Swami Vivekananda (1863-1902): the nine-volume English Complete Works and the ten-volume Bengali Vani o Rachana, two related but non-parallel corpora of about 15 million characters. Four layers are released: (i) a structure-preserving parse (32,694 paragraphs, 168,842 sentences) with per-paragraph anchors deep-linking into the published editions; (ii) a cross-lingual concept graph of 8,362 language-agnostic concepts with 87,518 relation-typed paragraph-concept and 55,872 concept-concept edges, in which canonical English labels act as a string-equality key linking Bengali and English passages with no parallel data; (iii) a bilingual alias inventory of 60,850 surface forms (30,053 English, 30,797 Bengali); and (iv) a human-annotated set of 200 paragraph-concept edges judged by three annotators, released with all per-annotator judgments. We report known-item cross-lingual retrieval over 194 verified rendered lecture pairs (Recall@10 0.86 in both directions), a 30-question audit of citation integrity and modern-question bridging, and a human study placing concept-extraction precision at 0.60 under strict two-annotator consensus (Cohen's kappa = 0.61). The extractor's confidence weight is calibrated: restricting to weight >= 0.8 raises precision to 0.71 while retaining 98% of concept-bearing paragraphs. Precision is markedly lower in Bengali than English (0.54 vs 0.68), locating the weakness in exactly the half that cross-lingual access depends on. The design transfers to other multilingual classical corpora. --- ### SAGG: Sample-Adaptive Gradient Gating for Robust Multimodal Learning under Heterogeneous Corruption Authors: Wentao Zhang, Yifan Zhu, Yutong Zhang, Wentao Mo Published: 2026-07-30 Categories: cs.LG, cs.MM arXiv: https://arxiv.org/abs/2609.20302v1 Multimodal gradient balancing methods modulate encoder gradients with a shared scalar per modality, implicitly assuming that corruption is uniform across the training batch. In practice, corruption is sample-heterogeneous: within a single mini-batch, different samples may have different modalities corrupted. We prove that under this heterogeneous corruption model, any batch-level sample-agnostic linear estimator with a shared modulation parameter incurs an irreducible bias with respect to the clean-data gradient, and that sample-level all-or-nothing gating is the unique unbiased strategy within a natural distribution-free estimator class. Motivated by this result, we propose Sample-Adaptive Gradient Gating (SAGG), which makes a binary retain-or-discard decision per sample via an online feature-norm quality test and incorporates a truncation mechanism for variance control. We prove that SAGG-based SGD converges at the standard O(1/sqrt(T)) rate to stationary points of the clean loss without a corruption-dependent error floor, and derive a certified robustness radius for the independent-encoder architecture that connects per-modality Lipschitz constants to the classification margin. Experiments on Kinetics-Sounds and UCF-101 under Gaussian noise injection, partial modality missing, and natural contribution imbalance show that SAGG consistently outperforms ten existing methods, with the largest gains in high-corruption regimes where batch-level bias is most severe. --- ### Not All Layers Are Equal: Dynamic Layer Routing for Reliable CLIP OOD Detection Authors: Ignacio M. De la Jara, Cristian Rodriguez-Opazo, Damith Ranasinghe Published: 2026-07-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.20299v1 Information aggregation across model layers are revealed to improve OOD detection. In contrast to crafting a method for layer-wise information aggregation in recent work, we investigate if layer selection is a learnable problem. In other words, we transpose the question from how to fuse layers to one asking which layers to trust for an input. Using a generalizable, weak, out of distribution context crafting approach for supervision, shown to be more effective than state of the art methods' mechanisms, we formulate learning a lightweight router to select a sparse, final-layer-anchored expert over CLIP's layer depth for OOD detection. Across three diverse benchmarks we demonstrate our learnable routing method dubbed Voyager improves OOD detection. On ImageNet-1K, Voyager achieves an average FPR@95 of 18.86, outperforming the strongest, comparable, prompt-learning method by 8.8 points. These gains persist across multiple supervision sources, including those used by existing state-of-the-art prompt-learning methods, demonstrating that, whilst our weak OOD supervision context is highly effective, the key advantage is realized from the learnable router component rather than the supervision source. Significantly, Voyager is highly practical; router learning takes approximately two minutes using less than 1 GB of memory, making it approximately 20x more efficient than current prompt-learning approaches. Anonymized Code: https://anonymous.4open.science/r/Voyager/ --- ### Queries Knew More Than We Thought: Uncovering Latent Knowledge in Segmentation Models Authors: Ignacio M. De la Jara, Cristian Rodriguez-Opazo, Damith Ranasinghe Published: 2026-07-30 Categories: cs.CV arXiv: https://arxiv.org/abs/2609.20283v1 Modern segmenters often fail after the expensive computation has already been done: a useful mask is present among the model's query-conditioned candidates, but the deployed selection rule does not expose it. We study this output-selection bottleneck in frozen DETR-family models. A ground-truth-only oracle first shows substantial hidden headroom in already-computed mask proposals. This raises a simple question: How can we better use the masks a segmenter has already computed but does not expose? We then ask whether that headroom can be recovered without adding queries, generating new masks, rerunning the backbone, or updating weights. HYDRA is a small selector trained only on cached frozen outputs. At inference time, it scores the cached candidates against an explicit keep-baseline option and acts only when a held-out calibrated margin indicates the selected candidate is sufficiently better. Trained on training-split caches and calibrated on held-out data, HYDRA improves Mask2Former, MaskDINO, and OneFormer by up to +7.41 dataset mIoU points on ADE20k and COCO, and improves SAM 3 by +9.4 class-macro prompt-IoU points on average across eight domains while preserving useful predictions through calibration. Paired LoRA controls show that lightweight weight adaptation does not remove the bottleneck: exposed predictions are often flat or worse, while routing over the adapted candidates still recovers accuracy. Finally, we connect the effect to query specialization under bipartite matching and verify it in a controlled TinyDETR study. These results show that frozen segmenters should be evaluated not only by the masks they expose, but also by the useful candidates they suppress. --- ### GroupKV: Hierarchical KV Cache Management for Long-Context Diffusion LLM Inference Authors: Jinhao Wang, Zhexin Hu, Kangjie Zhou, Xin Zhou, Fangfang Liu Published: 2026-07-30 Categories: cs.OS, cs.DC, cs.LG arXiv: https://arxiv.org/abs/2609.17573v1 Diffusion large language models (dLLMs) are emerging as a promising generative paradigm that complements autoregressive decoding. In long-context settings, KV cache bloat and offloading transfer overhead have become primary bottlenecks in inference systems. Meanwhile, the periodic full-sequence recomputation and localized token updates in dLLMs make the KV lifecycle substantially more dynamic, complicating cache management and prefetch scheduling while making heavyweight token-level indexing or clustering schemes harder to amortize effectively during decoding. To address these challenges, we present \textsc{GroupKV}, a lightweight hierarchical KV cache management system for long-context dLLM inference. We observe that under block-wise decoding, tokens within the same generation block tend to access highly overlapping and spatially concentrated context regions, making group-level sparse selection effective. Building on this observation, \textsc{GroupKV} partitions the context into contiguous groups and performs coarse-to-fine sparse selection. \textsc{GroupKV} further exploits cross-layer consistency to enable predictive prefetching, and incorporates a staleness correction mechanism to maintain cache coherence under dynamic KV updates. Additionally, \textsc{GroupKV} adopts streaming prefill to reduce peak memory consumption during prefilling. Experiments show that \textsc{GroupKV} extends the maximum serviceable context length by up to $48.00\times$ under constrained GPU memory, improves end-to-end inference performance by up to $3.73\times$ in offload-based long-context settings, and maintains competitive task accuracy. --- ### Speaker or Language? Explaining Variance in Charismatic Prosody Across Luxembourgish and French Authors: Nina Hosseini-Kivanani, Nafiseh Taghva, Peter Gilles, Oliver Niebuhr Published: 2026-07-30 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16275v1 Charismatic speech is shaped by language and speaking style, yet their relative contribution in bilingual public speaking remains unclear. We analyzed spontaneous speeches of 10 politicians who address audiences in Luxembourgish and French, in highly comparable communicative contexts across languages. From 400 utterances, we extracted 41 acoustic-prosodic features linked to vocal charisma and fitted mixed-effects models to separate speaker- and language-related variance. Speaker identity accounted for most variance, whereas language explained less, but still showed systematic differences: French productions showed higher shimmer and phrase-final F0, indicative of a polite, respectful voice, while Luxembourgish productions exhibited stronger mid-frequency spectral energy, suggesting a more vocally present profile. These patterns align with the sociolinguistic roles of Luxembourgish as an informal identity language and French as a high-prestige institutional variety. --- ### Speaker-Specific and Language-Dependent Temporal Organization in Bilingual Political Speech Authors: Nina Hosseini-Kivanani, Nafiseh Taghva, Peter Gilles, Oliver Niebuhr Published: 2026-07-30 Categories: cs.CL arXiv: https://arxiv.org/abs/2609.16274v1 Speech rhythm helps structure persuasive speech, but most empirical work examines monolingual English. This study asks how politicians organize timing when speaking Luxembourgish and French. We analyze 400 sentences from ten politicians, annotated for segments and pauses. We compute rhythm metrics, including means, variability, and pairwise variability indices for consonants and vowels. We quantify speaker and language contributions and test within-speaker language effects with paired t-tests. Results show that consonant-based metrics retain speaker-specific signatures, whereas vowel-based metrics are largely driven by language choice. French tokens display longer and more variable vowels and vocalic intervals, while consonant timing differences are smaller. No robust language by gender interactions emerge. These findings show that language choice systematically reorganizes rhythmic timing in bilingual public speech. --- ### Models as Governed Interfaces for AI-Native MBSE: Read-Side Adequacy and Write-Side Admissibility Authors: Jason Gower, Michael J. de C. Henshaw, Siyuan Ji Published: 2026-07-30 Categories: cs.SE, cs.AI arXiv: https://arxiv.org/abs/2609.16252v2 Machine-readable models such as SysML v2 are now programmatically accessible, and a growing body of work treats that access as the enabling condition for AI participation in systems engineering. Access is necessary, but not sufficient. The remaining work lies not in the modelling language but in the data architecture around it. An AI reader that queries a structurally complete model for a derivation still runs into absent derivation chains, untagged epistemic status, missing provenance, and evidence that the model cannot resolve. Faced with these gaps, it does not abstain; it fills them from training data, a source that is neither verifiable nor governed. To make the case on a model that is exemplary by current practice rather than deficient, we probe the public Apollo 11 SysML v2 reconstruction. We name the missing property epistemic adequacy and offer it as a candidate data-architecture pattern in two halves. Read-side adequacy lets derivation, status, and provenance answer a query rather than invite a guess; write-side admissibility gates an AI contribution before it enters the record. The property is broken down into five criteria. Four sit on the read side, evidenced by the case and convergent literature; the fifth sits on the participation side, advanced as a hypothesis this paper does not yet test. The architecture space runs from an inline metadata extension up to a substrate-native multi-model store, and over it, we propose the Governed-Query Architecture Framework, which governs agent participation through the viewpoint conventions that engineers already use. We commit the reframing to falsification: the epistemic layer counts as refuted if it cannot beat a retrieval-augmented baseline on the same model, tested first on the Apollo chain and then in an industrial pilot. --- ### Beyond Cultural Knowledge: Evaluating Arabic Cultural Appropriateness of Large Language Models Authors: Enes Altinisik, Hamdy Mubarak, Masoomali Fatehkia, Husrev_Taha_Sencar Husrev Taha Sencar Published: 2026-07-30 Categories: cs.CY, cs.CL arXiv: https://arxiv.org/abs/2609.16006v1 Large language models (LLMs) increasingly serve users whose expectations are shaped by their cultural context, yet most cultural evaluations test what a model knows rather than how it behaves when giving open-ended recommendations, opinions, and guidance. We introduce AraBehave: 1,623 culturally grounded, open-ended Arabic prompts with 29,214 cultural-appropriateness judgments from native speakers across several Arab regions, plus a scoring model whose predictions correlate strongly with human judgments on unseen systems (Pearson r=0.74). Evaluating three Arabic-centric and three frontier LLMs, we find that cultural appropriateness is not a single capability but decomposes into two largely independent components: normative stance and grounded cultural accuracy. The best general-purpose and best Arabic-centric models score identically (3.84 vs. 3.83 of 5) yet almost never fail for the same reason: general-purpose models exhibit strong factual grounding but a culturally inappropriate normative stance, being penalized for secular framing and false balance on culturally settled matters (28--33% of their low-score rationales), while the best Arabic-centric model adopts the expected stance but is penalized for fabricated hadith and misquoted verses (29%). Stance is cheap and fragile: one sentence of cultural instruction lifts Gemini to 4.57, above every Arabic-specialized model. Conversely, a generic ``answer clearly and objectively'' prompt costs Allam-7B 0.68 points, while asking the same questions in English lowers scores for every model but one. Grounding instead tracks scale and Arabic alignment data, and disappears when culturally aware instruction tuning is replaced by a culture-neutral corpus. General safety benchmarks see none of this: they saturate above 89 while cultural scores span 2.71-3.84. We will release the benchmark, annotations, and the scoring model. --- ### MA-RAG: Multi-Agent Retrieval-Augmented Generation for Query-Driven Summarization of Longitudinal Parkinson's Disease Assessments Authors: Sana Alamgeera, Denise Goberta, Muhammad Irshad, Anne H. H. Ngu Published: 2026-07-30 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2608.28624v1 Accurate interpretation of single-visit and longitudinal clinical assessments for Parkinson's disease is time-consuming and often depends on specialist expertise. Although large language models (LLMs) can generate natural language summaries, they frequently lack domain-specific clinical grounding and struggle to produce factually correct and temporally consistent responses for structured longitudinal assessment data. To address these limitations, we propose MA-RAG, a query-driven multi-agent retrieval-augmented generation framework that decomposes clinical reasoning into domain-specialized agents, combines structured fact extraction, and synthesizes clinically grounded summaries through a final verification stage. The framework supports four clinical analysis tasks: single-session, trajectory, comparison, and cohort summarization. We evaluate MA-RAG using objective metrics, namely Fact Precision, Hallucination Rate, Temporal Fidelity, and Semantic Similarity, together with subjective evaluations conducted by clinical experts. Compared to Traditional, RAG-only, and Single-agent RAG baselines, MA-RAG substantially improves factual correctness, achieving up to a 122% relative increase in Fact Precision (from 0.436 to 0.990) and reducing the Hallucination Rate by up to 98% (from 0.564 to 0.010), while consistently receiving top ratings from clinical experts for organization and clinical usefulness. These results demonstrate that domain-specialized multi-agent reasoning enables reliable query-driven summarization of structured longitudinal clinical assessment data. --- ### Looking Again: Measuring Sycophancy in the Reasoning Chains of Multimodal Models Under Pressure Authors: Mahir Numayeer Islam, Gakuto Okuyama, Nikolaus Siauw, Shivank Garg, Madhur Panwar, Vasu Sharma Published: 2026-07-30 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2608.28623v1 Large multimodal reasoning models (LMRMs) are getting increasingly capable, primarily through generating explicit chain-of-thought reasoning before answering. In language models it has been observed that this performance often comes with sycophancy, the tendency of a model to agree with the user over the evidence. However, for LMRMs no reliable method to measure sycophancy yet exists. We bridge this gap by introducing a benchmark and dataset for evaluating LMRM sycophancy when confronted with a wrong answer from a user. Our benchmark pairs four visually grounded datasets spanning mathematical, clinical, temporal, and demographic reasoning with five pressure conditions in single-turn and multi-turn settings. We evaluate sycophancy in the final answer as well as its emergence within the reasoning chain. We find that sycophancy is prevalent under pressure, with Statement pressure eliciting the highest rates and Conviction the lowest for all models except Mistral-Small-4, and under multi-turn pressure reasoning-level sycophancy intensifies sharply in clinical visual judgement, reaching 95.7% for the most affected model. We further introduce a failure taxonomy separating reasoning-chain from answer-level sycophancy, and a complementary sentence-level taxonomy locating where in the chain drift first emerges. Our results show that sycophancy can corrupt the reasoning chain independently of the final answer, so answer-level evaluation alone is insufficient. ---