# Research Papers — 2026-03 · bundle page 82 of 243 > Full-text markdown of papers #4051–4100 of 12123 published in 2026-03, newest first. AI-readable mirrors of the full 12123-paper 2026-03 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-03/page/82 [← Previous bundle](https://aipod.dev/papers/archive/2026-03/page/81.md) [Next bundle →](https://aipod.dev/papers/archive/2026-03/page/83.md) ### Decorrelation, Diversity, and Emergent Intelligence: The Isomorphism Between Social Insect Colonies and Ensemble Machine Learning Authors: Ernest Fokoué, Gregory Babbitt, Yuval Levental Published: 2026-03-20 Categories: stat.ML, cs.LG arXiv: https://arxiv.org/abs/2603.20328v2 Social insect colonies and ensemble machine learning methods represent two of the most successful examples of decentralized information processing in nature and computation respectively. Here we develop a rigorous mathematical framework demonstrating that ant colony decision-making and random forest learning are isomorphic under a common formalism of \textbf{stochastic ensemble intelligence}. We show that the mechanisms by which genetically identical ants achieve functional differentiation -- through stochastic response to local cues and positive feedback -- map precisely onto the bootstrap aggregation and random feature subsampling that decorrelate decision trees. Using tools from Bayesian inference, multi-armed bandit theory, and statistical learning theory, we prove that both systems implement identical variance reduction strategies through decorrelation of identical units. We derive explicit mappings between ant recruitment rates and tree weightings, pheromone trail reinforcement and out-of-bag error estimation, and quorum sensing and prediction averaging. This isomorphism suggests that collective intelligence, whether biological or artificial, emerges from a universal principle: \textbf{randomized identical agents + diversity-enforcing mechanisms $\rightarrow$ emergent optimality}. --- ### Probing the Latent World: Emergent Discrete Symbols and Physical Structure in Latent Representations Authors: Liu hung ming Published: 2026-03-20 Categories: cs.LG, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2603.20327v1 Video world models trained with Joint Embedding Predictive Architectures (JEPA) acquire rich spatiotemporal representations by predicting masked regions in latent space rather than reconstructing pixels. This removes the visual verification pathway of generative models, creating a structural interpretability gap: the encoder has learned physical structure inaccessible in any inspectable form. Existing probing methods either operate in continuous space without a structured intermediate layer, or attach generative components whose parameters confound attribution of behavior to the encoder. We propose the AI Mother Tongue (AIM) framework as a passive quantization probe: a lightweight, vocabulary-free probe that converts V-JEPA 2 continuous latent vectors into discrete symbol sequences without task-specific supervision or modifying the encoder. Because the encoder is kept completely frozen, any symbolic structure in the AIM codebook is attributable entirely to V-JEPA 2 pre-trained representations -- not to the probe. We evaluate through category-contrast experiments on Kinetics-mini along three physical dimensions: grasp angle, object geometry, and motion temporal structure. AIM symbol distributions differ significantly across all three experiments (chi^2 p < 10^{-4}; MI 0.036--0.117 bits, NMI 1.2--3.9% of the 3-bit maximum; JSD up to 0.342; codebook active ratio 62.5%). The experiments reveal that V-JEPA 2 latent space is markedly compact: diverse action categories share a common representational core, with semantic differences encoded as graded distributional variations rather than categorical boundaries. These results establish Stage 1 of a four-stage roadmap toward an action-conditioned symbolic world model, demonstrating that structured symbolic manifolds are discoverable properties of frozen JEPA latent spaces. --- ### Prompt-Free Lightweight SAM Adaptation for Histopathology Nuclei Segmentation with Strong Cross-Dataset Generalization Authors: Muhammad Hassan Maqsood, Yanming Zhu, Alfred Lam, Getamesay Dagnaw, Xuefei Yin, Alan Wee-Chung Liew Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20326v1 Histopathology nuclei segmentation is crucial for quantitative tissue analysis and cancer diagnosis. Although existing segmentation methods have achieved strong performance, they are often computationally heavy and show limited generalization across datasets, which constrains their practical deployment. Recent SAM-based approaches have shown great potential in general and medical imaging, but typically rely on prompt guidance or complex decoders, making them less suitable for histopathology images with dense nuclei and heterogeneous appearances. We propose a prompt-free and lightweight SAM adaptation that leverages multi-level encoder features and residual decoding for accurate and efficient nuclei segmentation. The framework fine-tunes only LoRA modules within the frozen SAM encoder, requiring just 4.1M trainable parameters. Experiments on three benchmark datasets TNBC, MoNuSeg, and PanNuke demonstrate state-of-the-art performance and strong cross-dataset generalization, highlighting the effectiveness and practicality of the proposed framework for histopathology applications. --- ### DCG-Net: Dual Cross-Attention with Concept-Value Graph Reasoning for Interpretable Medical Diagnosis Authors: Getamesay Dagnaw, Xuefei Yin, Muhammad Hassan Maqsood, Yanming Zhu, Alan Wee-Chung Liew Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20325v1 Deep learning models have achieved strong performance in medical image analysis, but their internal decision processes remain difficult to interpret. Concept Bottleneck Models (CBMs) partially address this limitation by structuring predictions through human-interpretable clinical concepts. However, existing CBMs typically overlook the contextual dependencies among concepts. To address these issues, we propose an end-to-end interpretable framework \emph{DCG-Net} that integrates multimodal alignment with structured concept reasoning. DCG-Net introduces a Dual Cross-Attention module that replaces cosine similarity matching with bidirectional attention between visual tokens and canonicalized textual concept-value prototypes, enabling spatially localized evidence attribution. To capture the relational structure inherent to clinical concepts, we develop a Parametric Concept Graph initialized with Positive Pointwise Mutual Information priors and refined through sparsity-controlled message passing. This formulation models inter-concept dependencies in a manner consistent with clinical domain knowledge. Experiments on white blood cell morphology and skin lesion diagnosis demonstrate that DCG-Net achieves state-of-the-art classification performance while producing clinically interpretable diagnostic explanations. --- ### When Agents Disagree: The Selection Bottleneck in Multi-Agent LLM Pipelines Authors: Artem Maryanskyy, Dmitry Budnikov, Alibek T. Kaliyev Published: 2026-03-20 Categories: cs.MA, cs.AI arXiv: https://arxiv.org/abs/2603.20324v2 Multi-agent LLM pipelines produce contradictory evidence on whether team diversity improves output quality: heterogeneous Mixture-of-Agents teams outperform single models, yet homogeneous Self-MoA teams consistently win under synthesis-based aggregation. We propose a resolution by identifying the selection bottleneck -- a crossover threshold in aggregation quality that determines whether diversity helps or hurts. Under this model, we obtain a closed-form crossover threshold $s^*$ (Proposition 1) that separates the regimes where diversity helps and hurts. In a targeted experiment spanning 42 tasks across 7 categories ($N=210$), a diverse team with judge-based selection achieves a win rate of 0.810 against a single-model baseline, while a homogeneous team scores 0.512 -- near chance (Glass's $Δ= 2.07$). Judge-based selection outperforms MoA-style synthesis by $Δ_{\mathrm{WR}} = +0.631$ -- the synthesis approach is preferred over the baseline in zero of 42 tasks by the judge panel. A decoupled evaluation with independent judges confirms all directional findings (Spearman $ρ= 0.90$). Exploratory evidence suggests that including a weaker model improves performance while reducing cost ($p < 10^{-4}$, not pre-registered). Our results suggest that selector quality may be a more impactful design lever than generator diversity in single-round generate-then-select pipelines. --- ### NCSTR: Node-Centric Decoupled Spatio-Temporal Reasoning for Video-based Human Pose Estimation Authors: Quang Dang Huynh, Xuefei Yin, Andrew Busch, Hugo G. Espinosa, Alan Wee-Chung Liew, Matthew T. O. Worsey, Yanming Zhu Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20323v1 Video-based human pose estimation remains challenged by motion blur, occlusion, and complex spatiotemporal dynamics. Existing methods often rely on heatmaps or implicit spatio-temporal feature aggregation, which limits joint topology expressiveness and weakens cross-frame consistency. To address these problems, we propose a novel node-centric framework that explicitly integrates visual, temporal, and structural reasoning for accurate pose estimation. First, we design a visuo-temporal velocity-based joint embedding that fuses sub-pixel joint cues and inter-frame motion to build appearance- and motion-aware representations. Then, we introduce an attention-driven pose-query encoder, which applies attention over joint-wise heatmaps and frame-wise features to map the joint representations into a pose-aware node space, generating image-conditioned joint-aware node embeddings. Building upon these node embeddings, we propose a dual-branch decoupled spatio-temporal attention graph that models temporal propagation and spatial constraint reasoning in specialized local and global branches. Finally, a node-space expert fusion module is proposed to adaptively fuse the complementary outputs from both branches, integrating local and global cues for final joint predictions. Extensive experiments on three widely used video pose benchmarks demonstrate that our method outperforms state-of-the-art methods. The results highlight the value of explicit node-centric reasoning, offering a new perspective for advancing video-based human pose estimation. --- ### MME-CoF-Pro: Evaluating Reasoning Coherence in Video Generative Models with Text and Visual Hints Authors: Yu Qi, Xinyi Xu, Ziyu Guo, Siyuan Ma, Renrui Zhang, Xinyan Chen, Ruichuan An, Ruofan Xing, Jiayi Zhang, Haojie Huang, Pheng-Ann Heng, Jonathan Tremblay et al. Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20194v1 Video generative models show emerging reasoning behaviors. It is essential to ensure that generated events remain causally consistent across frames for reliable deployment, a property we define as reasoning coherence. To bridge the gap in literature for missing reasoning coherence evaluation, we propose MME-CoF-Pro, a comprehensive video reasoning benchmark to assess reasoning coherence in video models. Specifically, MME-CoF-Pro contains 303 samples across 16 categories, ranging from visual logical to scientific reasoning. It introduces Reasoning Score as evaluation metric for assessing process-level necessary intermediate reasoning steps, and includes three evaluation settings, (a) no hint (b) text hint and (c) visual hint, enabling a controlled investigation into the underlying mechanisms of reasoning hint guidance. Evaluation results in 7 open and closed-source video models reveals insights including: (1) Video generative models exhibit weak reasoning coherence, decoupled from generation quality. (2) Text hints boost apparent correctness but often cause inconsistency and hallucinated reasoning (3) Visual hints benefit structured perceptual tasks but struggle with fine-grained perception. Website: https://video-reasoning-coherence.github.io/ --- ### From Masks to Pixels and Meaning: A New Taxonomy, Benchmark, and Metrics for VLM Image Tampering Authors: Xinyi Shang, Yi Tang, Jiacheng Cui, Ahmed Elhagry, Salwa K. Al Khatib, Sondos Mahmoud Bsharat, Jiacheng Liu, Xiaohan Zhao, Jing-Hao Xue, Hao Li, Salman Khan, Zhiqiang Shen Published: 2026-03-20 Categories: cs.CV, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.20193v1 Existing tampering detection benchmarks largely rely on object masks, which severely misalign with the true edit signal: many pixels inside a mask are untouched or only trivially modified, while subtle yet consequential edits outside the mask are treated as natural. We reformulate VLM image tampering from coarse region labels to a pixel-grounded, meaning and language-aware task. First, we introduce a taxonomy spanning edit primitives (replace/remove/splice/inpaint/attribute/colorization, etc.) and their semantic class of tampered object, linking low-level changes to high-level understanding. Second, we release a new benchmark with per-pixel tamper maps and paired category supervision to evaluate detection and classification within a unified protocol. Third, we propose a training framework and evaluation metrics that quantify pixel-level correctness with localization to assess confidence or prediction on true edit intensity, and further measure tamper meaning understanding via semantics-aware classification and natural language descriptions for the predicted regions. We also re-evaluate the existing strong segmentation/localization baselines on recent strong tamper detectors and reveal substantial over- and under-scoring using mask-only metrics, and expose failure modes on micro-edits and off-mask changes. Our framework advances the field from masks to pixels, meanings and language descriptions, establishing a rigorous standard for tamper localization, semantic classification and description. Code and benchmark data are available at https://github.com/VILA-Lab/PIXAR. --- ### LumosX: Relate Any Identities with Their Attributes for Personalized Video Generation Authors: Jiazheng Xing, Fei Du, Hangjie Yuan, Pengwei Liu, Hongbin Xu, Hai Ci, Ruigang Niu, Weihua Chen, Fan Wang, Yong Liu Published: 2026-03-20 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2603.20192v1 Recent advances in diffusion models have significantly improved text-to-video generation, enabling personalized content creation with fine-grained control over both foreground and background elements. However, precise face-attribute alignment across subjects remains challenging, as existing methods lack explicit mechanisms to ensure intra-group consistency. Addressing this gap requires both explicit modeling strategies and face-attribute-aware data resources. We therefore propose LumosX, a framework that advances both data and model design. On the data side, a tailored collection pipeline orchestrates captions and visual cues from independent videos, while multimodal large language models (MLLMs) infer and assign subject-specific dependencies. These extracted relational priors impose a finer-grained structure that amplifies the expressive control of personalized video generation and enables the construction of a comprehensive benchmark. On the modeling side, Relational Self-Attention and Relational Cross-Attention intertwine position-aware embeddings with refined attention dynamics to inscribe explicit subject-attribute dependencies, enforcing disciplined intra-group cohesion and amplifying the separation between distinct subject clusters. Comprehensive evaluations on our benchmark demonstrate that LumosX achieves state-of-the-art performance in fine-grained, identity-consistent, and semantically aligned personalized multi-subject video generation. Code and models are available at https://jiazheng-xing.github.io/lumosx-home/. --- ### Deterministic Mode Proposals: An Efficient Alternative to Generative Sampling for Ambiguous Segmentation Authors: Sebastian Gerard, Josephine Sullivan Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20191v1 Many segmentation tasks, such as medical image segmentation or future state prediction, are inherently ambiguous, meaning that multiple predictions are equally correct. Current methods typically rely on generative models to capture this uncertainty. However, identifying the underlying modes of the distribution with these methods is computationally expensive, requiring large numbers of samples and post-hoc clustering. In this paper, we shift the focus from stochastic sampling to the direct generation of likely outcomes. We introduce mode proposal models, a deterministic framework that efficiently produces a fixed-size set of proposal masks in a single forward pass. To handle superfluous proposals, we adapt a confidence mechanism, traditionally used in object detection, to the high-dimensional space of segmentation masks. Our approach significantly reduces inference time while achieving higher ground-truth coverage than existing generative models. Furthermore, we demonstrate that our model can be trained without knowing the full distribution of outcomes, making it applicable to real-world datasets. Finally, we show that by decomposing the velocity field of a pre-trained flow model, we can efficiently estimate prior mode probabilities for our proposals. --- ### CoVR-R:Reason-Aware Composed Video Retrieval Authors: Omkar Thawakar, Dmitry Demidov, Vaishnav Potlapalli, Sai Prasanna Teja Reddy Bogireddy, Viswanatha Reddy Gajjala, Alaa Mostafa Lasheen, Rao Muhammad Anwer, Fahad Khan Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20190v2 Composed Video Retrieval (CoVR) aims to find a target video given a reference video and a textual modification. Prior work assumes the modification text fully specifies the visual changes, overlooking after-effects and implicit consequences (e.g., motion, state transitions, viewpoint or duration cues) that emerge from the edit. We argue that successful CoVR requires reasoning about these after-effects. We introduce a reasoning-first, zero-shot approach that leverages large multimodal models to (i) infer causal and temporal consequences implied by the edit, and (ii) align the resulting reasoned queries to candidate videos without task-specific finetuning. To evaluate reasoning in CoVR, we also propose CoVR-Reason, a benchmark that pairs each (reference, edit, target) triplet with structured internal reasoning traces and challenging distractors that require predicting after-effects rather than keyword matching. Experiments show that our zero-shot method outperforms strong retrieval baselines on recall at K and particularly excels on implicit-effect subsets. Our automatic and human analysis confirm higher step consistency and effect factuality in our retrieved results. Our findings show that incorporating reasoning into general-purpose multimodal models enables effective CoVR by explicitly accounting for causal and temporal after-effects. This reduces dependence on task-specific supervision, improves generalization to challenging implicit-effect cases, and enhances interpretability of retrieval outcomes. These results point toward a scalable and principled framework for explainable video search. The model, code, and benchmark are available at https://github.com/mbzuai-oryx/CoVR-R. --- ### Learning Sampled-data Control for Swarms via MeanFlow Authors: Anqi Dong, Yongxin Chen, Karl H. Johansson, Johan Karlsson Published: 2026-03-20 Categories: cs.LG, cs.MA, cs.RO, eess.SY arXiv: https://arxiv.org/abs/2603.20189v2 Steering large-scale swarms with only limited control updates is often needed due to communication or computational constraints, yet most learning-based approaches do not account for this and instead model instantaneous velocity fields. As a result, the natural object for decision making is a finite-window control quantity rather than an infinitesimal one. To address this gap, we consider the recent machine learning framework MeanFlow and generalize it to the setting with general linear dynamic systems. This results in a new sampled-data learning framework that operates directly in control space and that can be applied for swarm steering. To this end, we learn the finite-horizon coefficient that parameterizes the minimum-energy control applied over each interval, and derive a differential identity that connects this quantity to a local bridge-induced supervision signal. This identity leads to a simple stop-gradient regression objective, allowing the interval coefficient field to be learned efficiently from bridge samples. The learned policy is deployed through sampled-data updates, guaranteeing that the resulting controller exactly respects the prescribed linear time-invariant dynamics and actuation channel. The resulting method enables few-step swarm steering at scale, while remaining consistent with the finite-window actuation structure of the underlying control system. --- ### Wildfire Spread Scenarios: Increasing Sample Diversity of Segmentation Diffusion Models with Training-Free Methods Authors: Sebastian Gerard, Josephine Sullivan Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20188v1 Predicting future states in uncertain environments, such as wildfire spread, medical diagnosis, or autonomous driving, requires models that can consider multiple plausible outcomes. While diffusion models can effectively learn such multi-modal distributions, naively sampling from these models is computationally inefficient, potentially requiring hundreds of samples to find low-probability modes that may still be operationally relevant. In this work, we address the challenge of sample-efficient ambiguous segmentation by evaluating several training-free sampling methods that encourage diverse predictions. We adapt two techniques, particle guidance and SPELL, originally designed for the generation of diverse natural images, to discrete segmentation tasks, and additionally propose a simple clustering-based technique. We validate these approaches on the LIDC medical dataset, a modified version of the Cityscapes dataset, and MMFire, a new simulation-based wildfire spread dataset introduced in this paper. Compared to naive sampling, these approaches increase the HM IoU* metric by up to 7.5% on MMFire and 16.4% on Cityscapes, demonstrating that training-free methods can be used to efficiently increase the sample diversity of segmentation diffusion models with little cost to image quality and runtime. Code and dataset: https://github.com/SebastianGer/wildfire-spread-scenarios --- ### NOUS: Video-Driven 3D Human Reaction Generation via Observation-Reaction Mutual Steering Authors: Yuan Zhou, Luanyuan Dai, Yongzhi Li, Shijie Hao, Xingyu Zhu, Yi Tan, Qingshan Xu, Beier Zhu, Richang Hong, Hanwang Zhang Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20187v3 Video-driven 3D human reaction generation aims to synthesize 3D human motion in response to the action observed in a video, playing an important role in interactive multimedia systems and embodied agents. Yet reaction motions generated by current methods often fail to match what the observed video calls for. We observe that one factor behind this failure is relational distortion in the correspondence between visual observations and reactions: videos lying close in the visual space may correspond to entirely different motions in the reaction space, which misleads the model into generating reactions inconsistent with the conditioning video. This motivates us to propose a new observatioN-reactiOn mUtual Steering (\texttt{NOUS}) framework that enables mutual steering between the video and motion modalities. It first performs Motion Feedback Steering (MFS), equipping the frozen pretrained video encoder with a lightweight rectification modulator and training the modulator with a relational margin loss that pulls each video embedding toward the motion prototype of its own category and away from those of other categories. In this way, the misaligned correspondence between visual observations and reactions can be calibrated. \texttt{NOUS} then applies Observation-Guided Refinement (OGR), which in turn exploits the rectified observations to further refine the generated reactions and enhance their quality. The results on the ViMo dataset demonstrate that \texttt{NOUS} improves the quality of reaction motion while incurring negligible computational overhead at inference. Also, \texttt{NOUS} yields consistent gains across four pretrained video encoders, showing its good compatibility. --- ### MuSteerNet: Human Reaction Generation from Videos via Observation-Reaction Mutual Steering Authors: Yuan Zhou, Yongzhi Li, Yanqi Dai, Xingyu Zhu, Yi Tan, Qingshan Xu, Beier Zhu, Richang Hong, Hanwang Zhang Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20187v2 Video-driven human reaction generation aims to synthesize 3D human motions that directly react to observed video sequences, which is crucial for building human-like interactive AI systems. However, existing methods often fail to effectively leverage video inputs to steer human reaction synthesis, resulting in reaction motions that are mismatched with the content of video sequences. We reveal that this limitation arises from a severe relational distortion between visual observations and reaction types. In light of this, we propose MuSteerNet, a simple yet effective framework that generates 3D human reactions from videos via observation-reaction mutual steering. Specifically, we first propose a Prototype Feedback Steering mechanism to mitigate relational distortion by refining visual observations with a gated delta-rectification modulator and a relational margin constraint, guided by prototypical vectors learned from human reactions. We then introduce Dual-Coupled Reaction Refinement that fully leverages rectified visual cues to further steer the refinement of generated reaction motions, thereby effectively improving reaction quality and enabling MuSteerNet to achieve competitive performance. Extensive experiments and ablation studies validate the effectiveness of our method. Code coming soon: https://github.com/zhouyuan888888/MuSteerNet. --- ### Improving Image-to-Image Translation via a Rectified Flow Reformulation Authors: Satoshi Iizuka, Shun Okamoto, Kazuhiro Fukui Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20186v2 In this work, we propose Image-to-Image Rectified Flow Reformulation (I2I-RFR), a practical plug-in reformulation that recasts standard I2I regression networks as continuous-time transport models. While pixel-wise I2I regression is simple, stable, and easy to adapt across tasks, it often over-smooths ill-posed and multimodal targets, whereas generative alternatives often require additional components, task-specific tuning, and more complex training and inference pipelines. Our method augments the backbone input by channel-wise concatenation with a noise-corrupted version of the ground-truth target and optimizes a simple t-reweighted pixel loss. This objective admits a rectified-flow interpretation via an induced velocity field, enabling ODE-based progressive refinement at inference time while largely preserving the standard supervised training pipeline. In most cases, adopting I2I-RFR requires only expanding the input channels, and inference can be performed with a few explicit solver steps (e.g., 3 steps) without distillation. Extensive experiments across multiple image-to-image translation and video restoration tasks show that I2I-RFR generally improves performance across a wide range of tasks and backbones, with particularly clear gains in perceptual quality and detail preservation. Overall, I2I-RFR provides a lightweight way to incorporate continuous-time refinement into conventional I2I models without requiring a heavy generative pipeline. --- ### Improving Image-to-Image Translation via a Rectified Flow Reformulation Authors: Satoshi Iizuka, Shun Okamoto, Kazuhiro Fukui Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20186v1 In this work, we propose Image-to-Image Rectified Flow Reformulation (I2I-RFR), a practical plug-in reformulation that recasts standard I2I regression networks as continuous-time transport models. While pixel-wise I2I regression is simple, stable, and easy to adapt across tasks, it often over-smooths ill-posed and multimodal targets, whereas generative alternatives often require additional components, task-specific tuning, and more complex training and inference pipelines. Our method augments the backbone input by channel-wise concatenation with a noise-corrupted version of the ground-truth target and optimizes a simple t-reweighted pixel loss. This objective admits a rectified-flow interpretation via an induced velocity field, enabling ODE-based progressive refinement at inference time while largely preserving the standard supervised training pipeline. In most cases, adopting I2I-RFR requires only expanding the input channels, and inference can be performed with a few explicit solver steps (e.g., 3 steps) without distillation. Extensive experiments across multiple image-to-image translation and video restoration tasks show that I2I-RFR generally improves performance across a wide range of tasks and backbones, with particularly clear gains in perceptual quality and detail preservation. Overall, I2I-RFR provides a lightweight way to incorporate continuous-time refinement into conventional I2I models without requiring a heavy generative pipeline. --- ### VideoSeek: Long-Horizon Video Agent with Tool-Guided Seeking Authors: Jingyang Lin, Jialian Wu, Jiang Liu, Ximeng Sun, Ze Wang, Xiaodong Yu, Jiebo Luo, Zicheng Liu, Emad Barsoum Published: 2026-03-20 Categories: cs.CV, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2603.20185v1 Video agentic models have advanced challenging video-language tasks. However, most agentic approaches still heavily rely on greedy parsing over densely sampled video frames, resulting in high computational cost. We present VideoSeek, a long-horizon video agent that leverages video logic flow to actively seek answer-critical evidence instead of exhaustively parsing the full video. This insight allows the model to use far fewer frames while maintaining, or even improving, its video understanding capability. VideoSeek operates in a think-act-observe loop with a well-designed toolkit for collecting multi-granular video observations. This design enables query-aware exploration over accumulated observations and supports practical video understanding and reasoning. Experiments on four challenging video understanding and reasoning benchmarks demonstrate that VideoSeek achieves strong accuracy while using far fewer frames than prior video agents and standalone LMMs. Notably, VideoSeek achieves a 10.2 absolute points improvement on LVBench over its base model, GPT-5, while using 93% fewer frames. Further analysis highlights the significance of leveraging video logic flow, strong reasoning capability, and the complementary roles of toolkit design. --- ### Kolmogorov-Arnold causal generative models Authors: Alejandro Almodóvar, Mar Elizo, Patricia A. Apellániz, Santiago Zazo, Juan Parras Published: 2026-03-20 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2603.20184v1 Causal generative models provide a principled framework for answering observational, interventional, and counterfactual queries from observational data. However, many deep causal models rely on highly expressive architectures with opaque mechanisms, limiting auditability in high-stakes domains. We propose KaCGM, a causal generative model for mixed-type tabular data where each structural equation is parameterized by a Kolmogorov--Arnold Network (KAN). This decomposition enables direct inspection of learned causal mechanisms, including symbolic approximations and visualization of parent--child relationships, while preserving query-agnostic generative semantics. We introduce a validation pipeline based on distributional matching and independence diagnostics of inferred exogenous variables, allowing assessment using observational data alone. Experiments on synthetic and semi-synthetic benchmarks show competitive performance against state-of-the-art methods. A real-world cardiovascular case study further demonstrates the extraction of simplified structural equations and interpretable causal effects. These results suggest that expressive causal generative modeling and functional transparency can be achieved jointly, supporting trustworthy deployment in tabular decision-making settings. Code: https://github.com/aalmodovares/kacgm --- ### Improving Generalization on Cybersecurity Tasks with Multi-Modal Contrastive Learning Authors: Jianan Huang, Rodolfo V. Valentim, Luca Vassio, Matteo Boffa, Marco Mellia, Idilio Drago, Dario Rossi Published: 2026-03-20 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2603.20181v1 The use of ML in cybersecurity has long been impaired by generalization issues: Models that work well in controlled scenarios fail to maintain performance in production. The root cause often lies in ML algorithms learning superficial patterns (shortcuts) rather than underlying cybersecurity concepts. We investigate contrastive multi-modal learning as a first step towards improving ML performance in cybersecurity tasks. We aim at transferring knowledge from data-rich modalities, such as text, to data-scarce modalities, such as payloads. We set up a case study on threat classification and propose a two-stage multi-modal contrastive learning framework that uses textual vulnerability descriptions to guide payload classification. First, we construct a semantically meaningful embedding space using contrastive learning on descriptions. Then, we align payloads to this space, transferring knowledge from text to payloads. We evaluate the approach on a large-scale private dataset and a synthetic benchmark built from public CVE descriptions and LLM-generated payloads. The methodology appears to reduce shortcut learning over baselines on both benchmarks. We release our synthetic benchmark and source code as open source. --- ### Adaptive Greedy Frame Selection for Long Video Understanding Authors: Yuning Huang, Xiaoyu Ji, Joseph Huang, Yichi Zhang, Fengqing Zhu Published: 2026-03-20 Categories: cs.CV, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2603.20180v2 Large vision--language models (VLMs) are increasingly applied to long-video question answering, yet inference is often bottlenecked by the number of input frames and resulting visual tokens. Naive sparse sampling can miss decisive moments, while purely relevance-driven selection frequently collapses onto near-duplicate frames and sacrifices coverage of temporally distant evidence. We propose a question-adaptive greedy frame selection method that jointly optimizes query relevance and semantic representativeness under a fixed frame budget. Our approach constructs a 1~FPS candidate pool (capped at 1000) with exact timestamp alignment, embeds candidates in two complementary spaces (SigLIP for question relevance and DINOv2 for semantic similarity), and selects frames by greedily maximizing a weighted sum of a modular relevance term and a facility-location coverage term. This objective is normalized, monotone, and submodular, yielding a standard (1-1/e) greedy approximation guarantee. To account for question-dependent trade-offs between relevance and coverage, we introduce four preset strategies and a lightweight text-only question-type classifier that routes each query to its best-performing preset. Experiments on MLVU show consistent accuracy gains over uniform sampling and a strong recent baseline across frame budgets, with the largest improvements under tight budgets. --- ### AI Agents Can Already Autonomously Perform Experimental High Energy Physics Authors: Eric A. Moreno, Samuel Bright-Thonney, Andrzej Novak, Dolores Garcia, Philip Harris Published: 2026-03-20 Categories: hep-ex, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.20179v3 Large language model-based AI agents are now able to autonomously execute substantial portions of a high energy physics (HEP) analysis pipeline with minimal expert-curated input. Given access to a HEP dataset, an execution framework, and a corpus of prior experimental literature, we find that Claude Code succeeds in automating all stages of a typical analysis: event selection, background estimation, uncertainty quantification, statistical inference, and paper drafting. We argue that the experimental HEP community is underestimating the current capabilities of these systems, and that most proposed agentic workflows are too narrowly scoped or scaffolded to specific analysis structures. We present a proof-of-concept framework, Just Furnish Context (JFC), that integrates autonomous analysis agents with literature-based knowledge retrieval and multi-agent review, and show that this is sufficient to plan, execute, and document a credible high energy physics analysis. We demonstrate this by conducting analyses on open data from ALEPH, DELPHI, and CMS to perform electroweak, QCD, and Higgs boson measurements. We present two of those results in a condensed short paper form -- a CMS Run1 Open Data $H\to τ^+τ^-$ to demonstrate performance on a well-established result, and the first Lund plane measurement on LEP data -- a genuinely novel result and, to our knowledge, the first produced autonomously by an AI agent. Rather than replacing physicists, these tools promise to offload the repetitive technical burden of analysis code development, freeing researchers to focus on physics insight, truly novel method development, and rigorous validation. Given these developments, we advocate for new strategies for how the community trains students, organizes analysis efforts, and allocates human expertise. --- ### LagerNVS: Latent Geometry for Fully Neural Real-time Novel View Synthesis Authors: Stanislaw Szymanowicz, Minghao Chen, Jianyuan Wang, Christian Rupprecht, Andrea Vedaldi Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20176v3 Recent work has shown that neural networks can perform 3D tasks such as Novel View Synthesis (NVS) without explicit 3D reconstruction. Even so, we argue that strong 3D inductive biases are still helpful in the design of such networks. We show this point by introducing LagerNVS, an encoder-decoder neural network for NVS that builds on `3D-aware' latent features. The encoder is initialized from a 3D reconstruction network pre-trained using explicit 3D supervision. This is paired with a lightweight decoder, and trained end-to-end with photometric losses. LagerNVS achieves state-of-the-art deterministic feed-forward Novel View Synthesis (including 31.4 PSNR on Re10k), with and without known cameras, renders in real time, generalizes to in-the-wild data, and can be paired with a diffusion decoder for generative extrapolation. --- ### TinyML Enhances CubeSat Mission Capabilities Authors: Luigi Capogrosso, Michele Magno Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20174v1 Earth observation (EO) missions traditionally rely on transmitting raw or minimally processed imagery from satellites to ground stations for computationally intensive analysis. This paradigm is infeasible for CubeSat systems due to stringent constraints on the onboard embedded processors, energy availability, and communication bandwidth. To overcome these limitations, the paper presents a TinyML-based Convolutional Neural Networks (ConvNets) model optimization and deployment pipeline for onboard image classification, enabling accurate, energy-efficient, and hardware-aware inference under CubeSat-class constraints. Our pipeline integrates structured iterative pruning, post-training INT8 quantization, and hardware-aware operator mapping to compress models and align them with the heterogeneous compute architecture of the STM32N6 microcontroller from STMicroelectronics. This Microcontroller Unit (MCU) integrates a novel Arm Cortex-M55 core and a Neural-ART Neural Processing Unit (NPU), providing a realistic proxy for CubeSat onboard computers. The paper evaluates the proposed approach on three EO benchmark datasets (i.e., EuroSAT, RS_C11, MEDIC) and four models (i.e., SqueezeNet, MobileNetV3, EfficientNet, MCUNetV1). We demonstrate an average reduction in RAM usage of 89.55% and Flash memory of 70.09% for the optimized models, significantly decreasing downlink bandwidth requirements while maintaining task-acceptable accuracy (with a drop ranging from 0.4 to 8.6 percentage points compared to the Float32 baseline). The energy consumption per inference ranges from 0.68 mJ to 6.45 mJ, with latency spanning from 3.22 ms to 30.38 ms. These results fully satisfy the stringent energy budgets and real-time constraints required for efficient onboard EO processing. --- ### Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluation Authors: Richard J. Young Published: 2026-03-20 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.20172v2 Recent work on chain-of-thought (CoT) faithfulness reports single aggregate numbers (e.g., DeepSeek-R1 acknowledges hints 39% of the time), implying that faithfulness is an objective, measurable property of a model. This paper provides evidence that it is not. Three classifiers (a regex-only detector, a regex-plus-LLM pipeline, and a Claude Sonnet 4 judge) are applied to 10,276 influenced reasoning traces from 12 open-weight models spanning 9 families and 7B to 1T parameters. On identical data, these classifiers produce faithfulness rates of 74.4%, 82.6%, and 69.7%. Per-model gaps range from 2.6 to 30.6 percentage points; all pairwise McNemar tests are significant (p < 0.001). The disagreements are systematic: Cohen's kappa ranges from 0.06 ("slight") for sycophancy hints to 0.42 ("moderate") for grader hints, and the asymmetry is pronounced: for sycophancy, 883 cases are classified as faithful by the pipeline but unfaithful by the Sonnet judge, while only 2 go the other direction. Classifier choice can also reverse model rankings: Qwen3.5-27B ranks 1st under the pipeline but 7th under Sonnet; OLMo-3.1-32B moves from 9th to 3rd. Different classifiers operationalize faithfulness at different levels of stringency (lexical mention versus epistemic dependence), yielding divergent measurements on the same behavior. These results indicate that published faithfulness numbers cannot be meaningfully compared across studies using different classifiers, and that future evaluations should report sensitivity ranges across multiple classification methodologies. --- ### Learning Dynamic Belief Graphs for Theory-of-mind Reasoning Authors: Ruxiao Chen, Xilei Zhao, Thomas J. Cova, Frank A. Drews, Susu Xu Published: 2026-03-20 Categories: cs.AI arXiv: https://arxiv.org/abs/2603.20170v2 Theory of Mind (ToM) reasoning with Large Language Models (LLMs) requires inferring how people's implicit, evolving beliefs shape what they seek and how they act under uncertainty -- especially in high-stakes settings such as disaster response, emergency medicine, and human-in-the-loop autonomy. Prior approaches either prompt LLMs directly or use latent-state models that treat beliefs as static and independent, often producing incoherent mental models over time and weak reasoning in dynamic contexts. We introduce a structured cognitive trajectory model for LLM-based ToM that represents mental state as a dynamic belief graph, jointly inferring latent beliefs, learning their time-varying dependencies, and linking belief evolution to information seeking and decisions. Our model contributes (i) a novel projection from textualized probabilistic statements to consistent probabilistic graphical model updates, (ii) an energy-based factor graph representation of belief interdependencies, and (iii) an ELBO-based objective that captures belief accumulation and delayed decisions. Across multiple real-world disaster evacuation datasets, our model significantly improves action prediction and recovers interpretable belief trajectories consistent with human reasoning, providing a principled module for augmenting LLMs with ToM in high-uncertainty environment. https://anonymous.4open.science/r/ICML_submission-6373/ --- ### Learning Dynamic Belief Graphs for Theory-of-mind Reasoning Authors: Ruxiao Chen, Xilei Zhao, Thomas J. Cova, Frank A. Drews, Susu Xu Published: 2026-03-20 Categories: cs.AI arXiv: https://arxiv.org/abs/2603.20170v1 Theory of Mind (ToM) reasoning with Large Language Models (LLMs) requires inferring how people's implicit, evolving beliefs shape what they seek and how they act under uncertainty -- especially in high-stakes settings such as disaster response, emergency medicine, and human-in-the-loop autonomy. Prior approaches either prompt LLMs directly or use latent-state models that treat beliefs as static and independent, often producing incoherent mental models over time and weak reasoning in dynamic contexts. We introduce a structured cognitive trajectory model for LLM-based ToM that represents mental state as a dynamic belief graph, jointly inferring latent beliefs, learning their time-varying dependencies, and linking belief evolution to information seeking and decisions. Our model contributes (i) a novel projection from textualized probabilistic statements to consistent probabilistic graphical model updates, (ii) an energy-based factor graph representation of belief interdependencies, and (iii) an ELBO-based objective that captures belief accumulation and delayed decisions. Across multiple real-world disaster evacuation datasets, our model significantly improves action prediction and recovers interpretable belief trajectories consistent with human reasoning, providing a principled module for augmenting LLMs with ToM in high-uncertainty environment. https://anonymous.4open.science/r/ICML_submission-6373/ --- ### EgoForge: Goal-Directed Egocentric World Simulator Authors: Yifan Shen, Jiateng Liu, Xinzhuo Li, Yuanzhe Liu, Bingxuan Li, Houze Yang, Wenqi Jia, Yijiang Li, Tianjiao Yu, James Matthew Rehg, Xu Cao, Ismini Lourentzou Published: 2026-03-20 Categories: cs.CV, cs.MM arXiv: https://arxiv.org/abs/2603.20169v1 Generative world models have shown promise for simulating dynamic environments, yet egocentric video remains challenging due to rapid viewpoint changes, frequent hand-object interactions, and goal-directed procedures whose evolution depends on latent human intent. Existing approaches either focus on hand-centric instructional synthesis with limited scene evolution, perform static view translation without modeling action dynamics, or rely on dense supervision, such as camera trajectories, long video prefixes, synchronized multicamera capture, etc. In this work, we introduce EgoForge, an egocentric goal-directed world simulator that generates coherent, first-person video rollouts from minimal static inputs: a single egocentric image, a high-level instruction, and an optional auxiliary exocentric view. To improve intent alignment and temporal consistency, we propose VideoDiffusionNFT, a trajectory-level reward-guided refinement that optimizes goal completion, temporal causality, scene consistency, and perceptual fidelity during diffusion sampling. Extensive experiments show EgoForge achieves consistent gains in semantic alignment, geometric stability, and motion fidelity over strong baselines, and robust performance in real-world smart-glasses experiments. --- ### The Robot's Inner Critic: Self-Refinement of Social Behaviors through VLM-based Replanning Authors: Jiyu Lim, Youngwoo Yoon, Kwanghyun Park Published: 2026-03-20 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2603.20164v1 Conventional robot social behavior generation has been limited in flexibility and autonomy, relying on predefined motions or human feedback. This study proposes CRISP (Critique-and-Replan for Interactive Social Presence), an autonomous framework where a robot critiques and replans its own actions by leveraging a Vision-Language Model (VLM) as a `human-like social critic.' CRISP integrates (1) extraction of movable joints and constraints by analyzing the robot's description file (e.g., MJCF), (2) generation of step-by-step behavior plans based on situational context, (3) generation of low-level joint control code by referencing visual information (joint range-of-motion visualizations), (4) VLM-based evaluation of social appropriateness and naturalness, including pinpointing erroneous steps, and (5) iterative refinement of behaviors through reward-based search. This approach is not tied to a specific robot API; it can generate subtly different, human-like motions on various platforms using only the robot's structure file. In a user study involving five different robot types and 20 scenarios, including mobile manipulators and humanoids, our proposed method achieved significantly higher preference and situational appropriateness ratings compared to previous methods. This research presents a general framework that minimizes human intervention while expanding the robot's autonomous interaction capabilities and cross-platform applicability. Detailed result videos and supplementary information regarding this work are available at: https://limjiyu99.github.io/inner-critic/ --- ### Evaluating Evidence Grounding Under User Pressure in Instruction-Tuned Language Models Authors: Sai Koneru, Elphin Joe, Christine Kirchhoff, Jian Wu, Sarah Rajtmajer Published: 2026-03-20 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.20162v1 In contested domains, instruction-tuned language models must balance user-alignment pressures against faithfulness to the in-context evidence. To evaluate this tension, we introduce a controlled epistemic-conflict framework grounded in the U.S. National Climate Assessment. We conduct fine-grained ablations over evidence composition and uncertainty cues across 19 instruction-tuned models spanning 0.27B to 32B parameters. Across neutral prompts, richer evidence generally improves evidence-consistent accuracy and ordinal scoring performance. Under user pressure, however, evidence does not reliably prevent user-aligned reversals in this controlled fixed-evidence setting. We report three primary failure modes. First, we identify a negative partial-evidence interaction, where adding epistemic nuance, specifically research gaps, is associated with increased susceptibility to sycophancy in families like Llama-3 and Gemma-3. Second, robustness scales non-monotonically: within some families, certain low-to-mid scale models are especially sensitive to adversarial user pressure. Third, models differ in distributional concentration under conflict: some instruction-tuned models maintain sharply peaked ordinal distributions under pressure, while others are substantially more dispersed; in scale-matched Qwen comparisons, reasoning-distilled variants (DeepSeek-R1-Qwen) exhibit consistently higher dispersion than their instruction-tuned counterparts. These findings suggest that, in a controlled fixed-evidence setting, providing richer in-context evidence alone offers no guarantee against user pressure without explicit training for epistemic integrity. --- ### Semantic Token Clustering for Efficient Uncertainty Quantification in Large Language Models Authors: Qi Cao, Andrew Gambardella, Takeshi Kojima, Yutaka Matsuo, Yusuke Iwasawa Published: 2026-03-20 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.20161v1 Large language models (LLMs) have demonstrated remarkable capabilities across diverse tasks. However, the truthfulness of their outputs is not guaranteed, and their tendency toward overconfidence further limits reliability. Uncertainty quantification offers a promising way to identify potentially unreliable outputs, but most existing methods rely on repeated sampling or auxiliary models, introducing substantial computational overhead. To address these limitations, we propose Semantic Token Clustering (STC), an efficient uncertainty quantification method that leverages the semantic information inherently encoded in LLMs. Specifically, we group tokens into semantically consistent clusters using embedding clustering and prefix matching, and quantify uncertainty based on the probability mass aggregated over the corresponding semantic cluster. Our approach requires only a single generation and does not depend on auxiliary models. Experimental results show that STC achieves performance comparable to state-of-the-art baselines while substantially reducing computational overhead. --- ### Beyond Single Tokens: Distilling Discrete Diffusion Models via Discrete MMD Authors: Emiel Hoogeboom, David Ruhe, Jonathan Heek, Thomas Mensink, Tim Salimans Published: 2026-03-20 Categories: cs.LG, cs.CV, stat.ML arXiv: https://arxiv.org/abs/2603.20155v1 It is currently difficult to distill discrete diffusion models. In contrast, continuous diffusion literature has many distillation approaches methods that can reduce sampling steps to a handful. Our method, Discrete Moment Matching Distillation (D-MMD), leverages ideas that have been highly successful in the continuous domain. Whereas previous discrete distillation methods collapse, D-MMD maintains high quality and diversity (given sufficient sampling steps). This is demonstrated on both text and image datasets. Moreover, the newly distilled generators can outperform their teachers. --- ### Design-OS: A Specification-Driven Framework for Engineering System Design with a Control-Systems Design Case Authors: H. Sinan Bank, Daniel R. Herber, Thomas H. Bradley Published: 2026-03-20 Categories: cs.CE, cs.AI, eess.SY arXiv: https://arxiv.org/abs/2603.20151v2 Engineering system design -- whether mechatronic, control, or embedded -- often proceeds in an ad hoc manner, with requirements left implicit and traceability from intent to parameters largely absent. Existing specification-driven and systematic design methods mostly target software, and AI-assisted tools tend to enter the workflow at solution generation rather than at problem framing. Human--AI collaboration in the design of physical systems remains underexplored. This paper presents Design-OS, a lightweight, specification-driven workflow for engineering system design organized in five stages: concept definition, literature survey, conceptual design, requirements definition, and design definition. Specifications serve as the shared contract between human designers and AI agents; each stage produces structured artifacts that maintain traceability and support agent-augmented execution. We position Design-OS relative to requirements-driven design, systematic design frameworks, and AI-assisted design pipelines, and demonstrate it on a control systems design case using two rotary inverted pendulum platforms -- an open-source SimpleFOC reaction wheel and a commercial Quanser Furuta pendulum -- showing how the same specification-driven workflow accommodates fundamentally different implementations. A blank template and the full design-case artifacts are shared in a public repository to support reproducibility and reuse. The workflow makes the design process visible and auditable, and extends specification-driven orchestration of AI from software to physical engineering system design. --- ### Enhancing Hyperspace Analogue to Language (HAL) Representations via Attention-Based Pooling for Text Classification Authors: Ali Sakour, Zoalfekar Sakour Published: 2026-03-20 Categories: cs.CL, cs.AI, cs.LG arXiv: https://arxiv.org/abs/2603.20149v1 The Hyperspace Analogue to Language (HAL) model relies on global word co-occurrence matrices to construct distributional semantic representations. While these representations capture lexical relationships effectively, aggregating them into sentence-level embeddings via standard mean pooling often results in information loss. Mean pooling assigns equal weight to all tokens, thereby diluting the impact of contextually salient words with uninformative structural tokens. In this paper, we address this limitation by integrating a learnable, temperature-scaled additive attention mechanism into the HAL representation pipeline. To mitigate the sparsity and high dimensionality of the raw co-occurrence matrices, we apply Truncated Singular Value Decomposition (SVD) to project the vectors into a dense latent space prior to the attention layer. We evaluate the proposed architecture on the IMDB sentiment analysis dataset. Empirical results demonstrate that the attention-based pooling approach achieves a test accuracy of 82.38%, yielding an absolute improvement of 6.74 percentage points over the traditional mean pooling baseline (75.64%). Furthermore, qualitative analysis of the attention weights indicates that the mechanism successfully suppresses stop-words and selectively attends to sentiment-bearing tokens, improving both classification performance and model interpretability. --- ### Can Large Multimodal Models Inspect Buildings? A Hierarchical Benchmark for Structural Pathology Reasoning Authors: Hui Zhong, Yichun Gao, Luyan Liu, Hai Yang, Wang Wang, Haowei Zhang, Xinhu Zheng Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20148v1 Automated building facade inspection is a critical component of urban resilience and smart city maintenance. Traditionally, this field has relied on specialized discriminative models (e.g., YOLO, Mask R-CNN) that excel at pixel-level localization but are constrained to passive perception and worse generization without the visual understandng to interpret structural topology. Large Multimodal Models (LMMs) promise a paradigm shift toward active reasoning, yet their application in such high-stakes engineering domains lacks rigorous evaluation standards. To bridge this gap, we introduce a human-in-the-loop semi-automated annotation framework, leveraging expert-proposal verification to unify 12 fragmented datasets into a standardized, hierarchical ontology. Building on this foundation, we present \textit{DefectBench}, the first multi-dimensional benchmark designed to interrogate LMMs beyond basic semantic recognition. \textit{DefectBench} evaluates 18 state-of-the-art (SOTA) LMMs across three escalating cognitive dimensions: Semantic Perception, Spatial Localization, and Generative Geometry Segmentation. Extensive experiments reveal that while current LMMs demonstrate exceptional topological awareness and semantic understanding (effectively diagnosing "what" and "how"), they exhibit significant deficiencies in metric localization precision ("where"). Crucially, however, we validate the viability of zero-shot generative segmentation, showing that general-purpose foundation models can rival specialized supervised networks without domain-specific training. This work provides both a rigorous benchmarking standard and a high-quality open-source database, establishing a new baseline for the advancement of autonomous AI agents in civil engineering. --- ### Synergistic Perception and Generative Recomposition: A Multi-Agent Orchestration for Expert-Level Building Inspection Authors: Hui Zhong, Yichun Gao, Luyan Liu, Xusen Guo, Zhaonian Kuang, Qiming Zhang, Xinhu Zheng Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20143v2 Building facade defect inspection is fundamental to structural health monitoring and sustainable urban maintenance, yet it remains a formidable challenge due to extreme geometric variability, low contrast against complex backgrounds, and the inherent complexity of composite defects (e.g., cracks co-occurring with spalling). Such characteristics lead to severe pixel imbalance and feature ambiguity, which, coupled with the critical scarcity of high-quality pixel-level annotations, hinder the generalization of existing detection and segmentation models. To address gaps, we propose \textit{FacadeFixer}, a unified multi-agent framework that treats defect perception as a collaborative reasoning task rather than isolated recognition. Specifically,\textit{FacadeFixer} orchestrates specialized agents for detection and segmentation to handle multi-type defect interference, working in tandem with a generative agent to enable semantic recomposition. This process decouples intricate defects from noisy backgrounds and realistically synthesizes them onto diverse clean textures, generating high-fidelity augmented data with precise expert-level masks. To support this, we introduce a comprehensive multi-task dataset covering six primary facade categories with pixel-level annotations. Extensive experiments demonstrate that \textit{FacadeFixer} significantly outperforms state-of-the-art (SOTA) baselines. Specifically, it excels in capturing pixel-level structural anomalies and highlights generative synthesis as a robust solution to data scarcity in infrastructure inspection. Our code and dataset will be made publicly available. --- ### Reasoning Gets Harder for LLMs Inside A Dialogue Authors: Ivan Kartáč, Mateusz Lango, Ondřej Dušek Published: 2026-03-20 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.20133v2 Large Language Models (LLMs) achieve strong performance on many reasoning benchmarks, yet these evaluations typically focus on isolated tasks that differ from real-world usage in task-oriented dialogue (TOD). In this setting, LLMs must perform reasoning inherently while generating text and adhering to instructions on role, format, and style. This mismatch raises concerns about whether benchmark performance accurately reflects models' reasoning robustness in TOD setting. We investigate how framing reasoning tasks within TOD affects LLM performance by introducing BOULDER, a new dynamic benchmark covering eight travel-related tasks that require arithmetic, spatial, and temporal reasoning with both commonsense and formal aspects. Each problem is presented in both isolated and dialogue-based variants, enabling controlled comparison while mitigating data contamination. Experiments on eight LLMs reveal a substantial and consistent performance gap between isolated and dialogue settings. Through ablations and qualitative analysis, we show that this gap is largely driven by the multi-turn nature of dialogue, with additional effects from role conditioning and tool-use requirements. Our results highlight the need to evaluate LLM reasoning in realistic interactive scenarios. --- ### Revisiting Gene Ontology Knowledge Discovery with Hierarchical Feature Selection and Virtual Study Group of AI Agents Authors: Cen Wan, Alex A. Freitas Published: 2026-03-20 Categories: cs.LG arXiv: https://arxiv.org/abs/2603.20132v1 Large language models have achieved great success in multiple challenging tasks, and their capacity can be further boosted by the emerging agentic AI techniques. This new computing paradigm has already started revolutionising the traditional scientific discovery pipelines. In this work, we propose a novel agentic AI-based knowledge discovery-oriented virtual study group that aims to extract meaningful ageing-related biological knowledge considering highly ageing-related Gene Ontology terms that are selected by hierarchical feature selection methods. We investigate the performance of the proposed agentic AI framework by considering four different model organisms' ageing-related Gene Ontology terms and validate the biological findings by reviewing existing research articles. It is found that the majority of the AI agent-generated scientific claims can be supported by existing literatures and the proposed internal mechanisms of the virtual study group also play an important role in the designed agentic AI-based knowledge discovery framework. --- ### An Agentic Multi-Agent Architecture for Cybersecurity Risk Management Authors: Ravish Gupta, Saket Kumar, Shreeya Sharma, Maulik Dang, Abhishek Aggarwal Published: 2026-03-20 Categories: eess.SY, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2603.20131v2 Getting a real cybersecurity risk assessment for a small organization is expensive -- a NIST CSF-aligned engagement runs $15,000 on the low end, takes weeks, and depends on practitioners who are genuinely scarce. Most small companies skip it entirely. We built a six-agent AI system where each agent handles one analytical stage: profiling the organization, mapping assets, analyzing threats, evaluating controls, scoring risks, and generating recommendations. Agents share a persistent context that grows as the assessment proceeds, so later agents build on what earlier ones concluded -- the mechanism that distinguishes this from standard sequential agent pipelines. We tested it on a 15-person HIPAA-covered healthcare company and compared outputs to independent assessments by three CISSP practitioners -- the system agreed with them 85% of the time on severity classifications, covered 92% of identified risks, and finished in under 15 minutes. We then ran 30 repeated single-agent assessments across five synthetic but sector-realistic organizational profiles in healthcare, fintech, manufacturing, retail, and SaaS, comparing a general-purpose Mistral-7B against a domain fine-tuned model. Both completed every run. The fine-tuned model flagged threats the baseline could not see at all: PHI exposure in healthcare, OT/IIoT vulnerabilities in manufacturing, platform-specific risks in retail. The full multi-agent pipeline, however, failed every one of 30 attempts on a Tesla T4 with its 4,096-token default context window -- context capacity, not model quality, turned out to be the binding constraint. --- ### Generalizable NGP-SR: Generalizable Neural Radiance Fields Super-Resolution via Neural Graph Primitives Authors: Wanqi Yuan, Omkar Sharad Mayekar, Connor Pennington, Nianyi Li Published: 2026-03-20 Categories: cs.CV arXiv: https://arxiv.org/abs/2603.20128v1 Neural Radiance Fields (NeRF) achieve photorealistic novel view synthesis but become costly when high-resolution (HR) rendering is required, as HR outputs demand dense sampling and higher-capacity models. Moreover, naively super-resolving per-view renderings in 2D often breaks multi-view consistency. We propose Generalizable NGP-SR, a 3D-aware super-resolution framework that reconstructs an HR radiance field directly from low-resolution (LR) posed images. Built on Neural Graphics Primitives (NGP), NGP-SR conditions radiance prediction on 3D coordinates and learned local texture tokens, enabling recovery of high-frequency details within the radiance field and producing view-consistent HR novel views without external HR references or post-hoc 2D upsampling. Importantly, our model is generalizable: once trained, it can be applied to unseen scenes and rendered from novel viewpoints without per-scene optimization. Experiments on multiple datasets show that NGP-SR consistently improves both reconstruction quality and runtime efficiency over prior NeRF-based super-resolution methods, offering a practical solution for scalable high-resolution novel view synthesis. --- ### Evolving Jailbreaks: Automated Multi-Objective Long-Tail Attacks on Large Language Models Authors: Wenjing Hong, Zhonghua Rong, Li Wang, Feng Chang, Jian Zhu, Ke Tang, Zexuan Zhu, Yew-Soon Ong Published: 2026-03-20 Categories: cs.CR, cs.AI arXiv: https://arxiv.org/abs/2603.20122v1 Large Language Models (LLMs) have been widely deployed, especially through free Web-based applications that expose them to diverse user-generated inputs, including those from long-tail distributions such as low-resource languages and encrypted private data. This open-ended exposure increases the risk of jailbreak attacks that undermine model safety alignment. While recent studies have shown that leveraging long-tail distributions can facilitate such jailbreaks, existing approaches largely rely on handcrafted rules, limiting the systematic evaluation of these security and privacy vulnerabilities. In this work, we present EvoJail, an automated framework for discovering long-tail distribution attacks via multi-objective evolutionary search. EvoJail formulates long-tail attack prompt generation as a multi-objective optimization problem that jointly maximizes attack effectiveness and minimizes output perplexity, and introduces a semantic-algorithmic solution representation to capture both high-level semantic intent and low-level structural transformations of encryption-decryption logic. Building upon this representation, EvoJail integrates LLM-assisted operators into a multi-objective evolutionary framework, enabling adaptive and semantically informed mutation and crossover for efficiently exploring a highly structured and open-ended search space. Extensive experiments demonstrate that EvoJail consistently discovers diverse and effective long-tail jailbreak strategies, achieving competitive performance with existing methods in both individual and ensemble level. --- ### Chain-of-Adaptation: Surgical Vision-Language Adaptation with Reinforcement Learning Authors: Jiajie Li, Chenhui Xu, Meihuan Liu, Jinjun Xiong Published: 2026-03-20 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2603.20116v1 Conventional fine-tuning on domain-specific datasets can inadvertently alter a model's pretrained multimodal priors, leading to reduced generalization. To address this, we propose Chain-of-Adaptation (CoA), an adaptation framework designed to integrate domain knowledge while maintaining the model's inherent reasoning and perceptual capabilities. CoA introduces a structured reasoning format that enhances domain alignment without sacrificing general multimodal competence by reinforcement learning. Experiments on standard surgical benchmarks, under both in-distribution and out-of-distribution settings, demonstrate that CoA achieves higher accuracy, stronger generalization, and more stable behavior than supervised fine-tuning. Furthermore, ablation studies confirm that CoA effectively preserves the model's core visual-language abilities, providing a reliable pathway for domain specialization in VLMs. --- ### Conditioning Protein Generation via Hopfield Pattern Multiplicity Authors: Jeffrey D. Varner Published: 2026-03-20 Categories: cs.LG, q-bio.BM, q-bio.QM arXiv: https://arxiv.org/abs/2603.20115v2 Small protein-family alignments often contain a subset of interest but not enough labeled data to train a conditional generator. We condition a training-free stochastic-attention sampler by adding one multiplicity ratio to its logits. Increasing this ratio shifts generation from the full family toward the designated subset. For unit-norm memories, the resulting Boltzmann distribution is exactly a Gaussian mixture whose component weights are set by the multiplicities. This result separates exact conditioning in latent space from losses caused by sampling, PCA reconstruction, and sequence decoding. Across five Pfam families, attention followed the analytic target, but recovery of single-residue markers depended on how well PCA separated the designated and background sequences. A matched weighted profile HMM reproduced these markers more directly, while stochastic attention gave lower ESM2 pseudo-perplexity in the Kunitz comparison. Using a curated set of 23 omega-conotoxin sequences as the target subset produced diverse sequences that preserved the cysteine scaffold and Tyr13 and shifted other residues toward the designated set. These sequences are candidates for experimental testing; they do not establish binding. --- ### Conditioning Protein Generation via Hopfield Pattern Multiplicity Authors: Jeffrey D. Varner Published: 2026-03-20 Categories: cs.LG, q-bio.BM, q-bio.QM arXiv: https://arxiv.org/abs/2603.20115v1 Protein sequence generation via stochastic attention produces plausible family members from small alignments without training, but treats all stored sequences equally and cannot direct generation toward a functional subset of interest. We show that a single scalar parameter, added as a bias to the sampler's attention logits, continuously shifts generation from the full family toward a user-specified subset, with no retraining and no change to the model architecture. A practitioner supplies a small set of sequences (for example, hits from a binding screen) and a multiplicity ratio that controls how strongly generation favors them. The method is agnostic to what the subset represents: binding, stability, specificity, or any other property. We find that the conditioning is exact at the level of the sampler's internal representation, but that the decoded sequence phenotype can fall short because the dimensionality reduction used to encode sequences does not always preserve the residue-level variation that defines the functional split. We term this discrepancy the calibration gap and show that it is predicted by a simple geometric measure of how well the encoding separates the functional subset from the rest of the family. Experiments on five Pfam families (Kunitz, SH3, WW, Homeobox, and Forkhead domains) confirm the monotonic relationship between separation and gap across a fourfold range of geometries. Applied to omega-conotoxin peptides targeting a calcium channel involved in pain signaling, curated seeding from 23 characterized binders produces over a thousand candidates that preserve the primary pharmacophore and all experimentally identified binding determinants. These results show that stochastic attention enables practitioners to expand a handful of experimentally characterized sequences into diverse candidate libraries without retraining a generative model. --- ### Current LLMs still cannot 'talk much' about grammar modules: Evidence from syntax Authors: Mohammed Q. Shormani, Yehia A. AlSohbani Published: 2026-03-20 Categories: cs.CL arXiv: https://arxiv.org/abs/2603.20114v4 We aim to examine the extent to which Large Language Models (LLMs) can 'talk much' about grammar modules, providing evidence from syntax core properties translated by ChatGPT into Arabic. We collected 44 terms from generative syntax previous works, including books and journal articles, as well as from our experience in the field. These terms were translated by humans, and then by ChatGPT-5. We then analyzed and compared both translations. We used an analytical and comparative approach in our analysis. Findings unveil that LLMs still cannot 'talk much' about the core syntax properties embedded in the terms under study involving several syntactic and semantic challenges: only 25% of ChatGPT translations were accurate, while 38.6% were inaccurate, and 36.4.% were partially correct, which we consider appropriate. Based on these findings, a set of actionable strategies were proposed, the most notable of which is a close collaboration between AI specialists and linguists to better LLMs' working mechanism for accurate or at least appropriate translation. --- ### Demonstration of Adapt4Me: An Uncertainty-Aware Authoring Environment for Personalizing Automatic Speech Recognition to Non-normative Speech Authors: Niclas Pokel, Yiming Zhao, Pehuén Moure, Yingqiang Gao, Roman Böhringer Published: 2026-03-20 Categories: cs.HC, cs.AI arXiv: https://arxiv.org/abs/2603.20112v1 Personalizing Automatic Speech Recognition (ASR) for non-normative speech remains challenging because data collection is labor-intensive and model training is technically complex. To address these limitations, we propose Adapt4Me, a web-based decentralized environment that operationalizes Bayesian active learning to enable end-to-end personalization without expert supervision. The app exposes data selection, adaptation, and validation to lay users through a three-stage human-in-the-loop workflow: (1) rapid profiling via greedy phoneme sampling to capture speaker-specific acoustics; (2) backend personalization using Variational Inference Low-Rank Adaptation (VI-LoRA) to enable fast, incremental updates; and (3) continuous improvement, where users guide model refinement by resolving visualized model uncertainty via low-friction top-k corrections. By making epistemic uncertainty explicit, Adapt4Me reframes data efficiency as an interactive design feature rather than a purely algorithmic concern. We show how this enables users to personalize robust ASR models, transforming them from passive data sources into active authors of their own assistive technology. --- ### Var-JEPA: A Variational Formulation of the Joint-Embedding Predictive Architecture - Bridging Predictive and Generative Self-Supervised Learning Authors: Moritz Gögl, Christopher Yau Published: 2026-03-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2603.20111v2 The Joint-Embedding Predictive Architecture (JEPA) is often seen as a non-generative alternative to likelihood-based self-supervised learning, emphasizing prediction in representation space rather than reconstruction in observation space. We argue that the resulting separation from probabilistic generative modeling is largely rhetorical rather than structural: the canonical JEPA design (coupled encoders with a context-to-target predictor) mirrors the variational posteriors and learned conditional priors obtained when variational inference is applied to a particular class of coupled latent-variable models, and standard JEPA can be viewed as a deterministic specialization in which regularization is imposed via architectural and training heuristics rather than an explicit likelihood. Building on this view, we derive the Variational JEPA (Var-JEPA), which makes the latent generative structure explicit by optimizing a single Evidence Lower Bound (ELBO). This yields meaningful representations without ad-hoc anti-collapse regularizers and allows principled uncertainty quantification in the latent space. We instantiate the framework for tabular data (Var-T-JEPA) and achieve strong representation learning and downstream performance, improving over T-JEPA across real-world tabular benchmarks while remaining competitive with strong raw-feature baselines. --- ### Var-JEPA: A Variational Formulation of the Joint-Embedding Predictive Architecture -- Bridging Predictive and Generative Self-Supervised Learning Authors: Moritz Gögl, Christopher Yau Published: 2026-03-20 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2603.20111v1 The Joint-Embedding Predictive Architecture (JEPA) is often seen as a non-generative alternative to likelihood-based self-supervised learning, emphasizing prediction in representation space rather than reconstruction in observation space. We argue that the resulting separation from probabilistic generative modeling is largely rhetorical rather than structural: the canonical JEPA design, coupled encoders with a context-to-target predictor, mirrors the variational posteriors and learned conditional priors obtained when variational inference is applied to a particular class of coupled latent-variable models, and standard JEPA can be viewed as a deterministic specialization in which regularization is imposed via architectural and training heuristics rather than an explicit likelihood. Building on this view, we derive the Variational JEPA (Var-JEPA), which makes the latent generative structure explicit by optimizing a single Evidence Lower Bound (ELBO). This yields meaningful representations without ad-hoc anti-collapse regularizers and allows principled uncertainty quantification in the latent space. We instantiate the framework for tabular data (Var-T-JEPA) and achieve strong representation learning and downstream performance, consistently improving over T-JEPA while remaining competitive with strong raw-feature baselines. --- ### GO-GenZip: Goal-Oriented Generative Sampling and Hybrid Compression Authors: Pietro Talli, Qi Liao, Alessandro Lieto, Parijat Bhattacharjee, Federico Chiariotti, Andrea Zanella Published: 2026-03-20 Categories: cs.LG, cs.IT arXiv: https://arxiv.org/abs/2603.20109v1 Current network data telemetry pipelines consist of massive streams of fine-grained Key Performance Indicators (KPIs) from multiple distributed sources towards central aggregators, making data storage, transmission, and real-time analysis increasingly unsustainable. This work presents a generative AI (GenAI)-driven sampling and hybrid compression framework that redesigns network telemetry from a goal-oriented perspective. Unlike conventional approaches that passively compress fully observed data, our approach jointly optimizes what to observe and how to encode it, guided by the relevance of information to downstream tasks. The framework integrates adaptive sampling policies, using adaptive masking techniques, with generative modeling to identify patterns and preserve critical features across temporal and spatial dimensions. The selectively acquired data are further processed through a hybrid compression scheme that combines traditional lossless coding with GenAI-driven, lossy compression. Experimental results on real network datasets demonstrate over 50$\%$ reductions in sampling and data transfer costs, while maintaining comparable reconstruction accuracy and goal-oriented analytical fidelity in downstream tasks. --- ### Trojan horse hunt in deep forecasting models: Insights from the European Space Agency competition Authors: Krzysztof Kotowski, Ramez Shendy, Jakub Nalepa, Agata Kaczmarek, Dawid Płudowski, Piotr Wilczyński, Artur Janicki, Przemysław Biecek, Ambros Marzetta, Atul Pande, Lalit Chandra Routhu, Swapnil Srivastava et al. Published: 2026-03-20 Categories: cs.LG, cs.CR arXiv: https://arxiv.org/abs/2603.20108v1 Forecasting plays a crucial role in modern safety-critical applications, such as space operations. However, the increasing use of deep forecasting models introduces a new security risk of trojan horse attacks, carried out by hiding a backdoor in the training data or directly in the model weights. Once implanted, the backdoor is activated by a specific trigger pattern at test time, causing the model to produce manipulated predictions. We focus on this issue in our \textit{Trojan Horse Hunt} data science competition, where more than 200 teams faced the task of identifying triggers hidden in deep forecasting models for spacecraft telemetry. We describe the novel task formulation, benchmark set, evaluation protocol, and best solutions from the competition. We further summarize key insights and research directions for effective identification of triggers in time series forecasting models. All materials are publicly available on the official competition webpage https://www.kaggle.com/competitions/trojan-horse-hunt-in-space. ---