# Research Papers — 2026-04 · bundle page 124 of 233 > Full-text markdown of papers #6151–6200 of 11612 published in 2026-04, newest first. AI-readable mirrors of the full 11612-paper 2026-04 archive (50 papers per file). Canonical page: https://aipod.dev/papers/archive/2026-04/page/124 [← Previous bundle](https://aipod.dev/papers/archive/2026-04/page/123.md) [Next bundle →](https://aipod.dev/papers/archive/2026-04/page/125.md) ### Token Encoding for Semantic Recovery Authors: Jingzhi Hu, Ouya Wang, Geoffrey Ye Li Published: 2026-04-14 Categories: eess.SP, cs.LG arXiv: https://arxiv.org/abs/2604.12931v2 In generative semantic communication, semantic tokens guide receiver-side generative models to synthesize high-dimensional content. In challenging network environments, however, frequent token erasure distorts the conveyed semantics beyond what receiver-side recovery can restore. In this paper, we propose a token encoding framework (TokCode) for robust semantic recovery, achieving erasure resilience by restructuring redundancy in the semantic domain. TokCode uses a lightweight adapter to recast a general-purpose large language model (LLM) at the transmitter into a token encoder, exploiting the LLM's pretrained semantic prior to avoid introducing a dedicated deep model. To optimize the adapter efficiently and make it applicable across diverse channels, we develop a channel-quality-aware distillation approach for token encoder training~(CADET). Using a differentiable sentence-level semantic surrogate, CADET tunes T5 foundation models into experts for distinct erasure rates and distills them into a single reconfigurable low-rank adapter, enabling subsequent reinforcement learning (RL) to start above the plateau where direct RL stalls. Simulation results on token-based generative image transmission show that TokCode improves the image-level similarity over the best-performing receiver-side recovery benchmark by 14.1%--22.4%, closing 71.9%--76.5% of its gap to the erasure-aware oracle encoding, when only 20% to 50% of the tokens survive. --- ### Token Encoding for Semantic Recovery Authors: Jingzhi Hu, Geoffrey Ye Li Published: 2026-04-14 Categories: eess.SP, cs.LG arXiv: https://arxiv.org/abs/2604.12931v1 Token-based semantic communication is promising for future wireless networks, as it can compact semantic tokens under very limited channel capacity. However, harsh wireless channels often cause missing tokens, leading to severe distortion that prevents reliable semantic recovery at the receiver. In this article, we propose a token encoding framework for robust semantic recovery (TokCode), which incurs no additional transmission overhead and supports plug-and-play deployment. For efficient token encoder optimization, we develop a sentence-semantic-guided foundation model adaptation algorithm (SFMA) that avoids costly end-to-end training. Based on simulation results on prompt-based generative image transmission, TokCode mitigates semantic distortion and can approach the performance upper-bound, even under harsh channels where 40% to 60% of tokens are randomly lost. --- ### Grasp in Gaussians: Fast Monocular Reconstruction of Dynamic Hand-Object Interactions Authors: Ayce Idil Aytekin, Xu Chen, Zhengyang Shen, Thabo Beeler, Helge Rhodin, Rishabh Dabral, Christian Theobalt Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12929v2 We present Grasp in Gaussians (GraG), a fast and robust method for reconstructing dynamic 3D hand-object interactions from a single monocular video. Unlike recent approaches that optimize heavy neural representations, our method focuses on tracking the hand and the object efficiently, once initialized from pretrained large models. Our key insight is that, given strong pretrained object and hand priors, accurate and temporally stable hand-object motion can be recovered using a compact Sum-of-Gaussians (SoG) representation, revived from classical tracking literature and integrated with generative Gaussian-based initializations. We initialize object pose and geometry using a video-adapted SAM3D pipeline, then convert the resulting dense Gaussian representation into a lightweight SoG via subsampling. This compact representation enables efficient and fast tracking while preserving geometric fidelity, with the pretrained priors and simple geometric/contact losses providing accurate hand-object placement. For the hand, we adopt a complementary strategy: starting from off-the-shelf monocular hand pose initialization, we refine hand motion using simple yet effective 2D joint, silhouette, depth, and contact losses, avoiding per-frame refinement of a detailed 3D hand appearance model while maintaining stable articulation. Extensive experiments on public benchmarks demonstrate that GraG reconstructs temporally coherent hand-object interactions on long sequences 4.4x-38.9x faster than prior work while preserving temporally coherent hand-object motion. --- ### Grasp in Gaussians: Fast Monocular Reconstruction of Dynamic Hand-Object Interactions Authors: Ayce Idil Aytekin, Xu Chen, Zhengyang Shen, Thabo Beeler, Helge Rhodin, Rishabh Dabral, Christian Theobalt Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12929v1 We present Grasp in Gaussians (GraG), a fast and robust method for reconstructing dynamic 3D hand-object interactions from a single monocular video. Unlike recent approaches that optimize heavy neural representations, our method focuses on tracking the hand and the object efficiently, once initialized from pretrained large models. Our key insight is that accurate and temporally stable hand-object motion can be recovered using a compact Sum-of-Gaussians (SoG) representation, revived from classical tracking literature and integrated with generative Gaussian-based initializations. We initialize object pose and geometry using a video-adapted SAM3D pipeline, then convert the resulting dense Gaussian representation into a lightweight SoG via subsampling. This compact representation enables efficient and fast tracking while preserving geometric fidelity. For the hand, we adopt a complementary strategy: starting from off-the-shelf monocular hand pose initialization, we refine hand motion using simple yet effective 2D joint and depth alignment losses, avoiding per-frame refinement of a detailed 3D hand appearance model while maintaining stable articulation. Extensive experiments on public benchmarks demonstrate that GraG reconstructs temporally coherent hand-object interactions on long sequences 6.4x faster than prior work while improving object reconstruction by 13.4% and reducing hand's per-joint position error by over 65%. --- ### MoshiRAG: Asynchronous Knowledge Retrieval for Full-Duplex Speech Language Models Authors: Chung-Ming Chien, Manu Orsini, Eugene Kharitonov, Neil Zeghidour, Karen Livescu, Alexandre Défossez Published: 2026-04-14 Categories: cs.CL, eess.AS arXiv: https://arxiv.org/abs/2604.12928v3 Speech-to-speech language models have recently emerged to enhance the naturalness of conversational AI. In particular, full-duplex models are distinguished by their real-time interactivity, including handling of pauses, interruptions, and backchannels. However, improving their factuality remains an open challenge. While scaling the model size could address this gap, it would make real-time inference prohibitively expensive. In this work, we propose MoshiRAG, a modular approach that combines a compact full-duplex interface with selective retrieval to access more powerful knowledge sources. Our asynchronous framework enables the model to identify knowledge-demanding queries and ground its responses in external information. By leveraging the natural temporal gap between response onset and the delivery of core information, the retrieval process can be completed while maintaining a natural conversation flow. With this approach, MoshiRAG achieves factuality comparable to the best publicly released non-duplex speech language models while preserving the interactivity inherent to full-duplex systems. Moreover, our flexible design supports plug-and-play retrieval methods without retraining and demonstrates strong performance on out-of-domain mathematical reasoning tasks. --- ### Pi-HOC: Pairwise 3D Human-Object Contact Estimation Authors: Sravan Chittupalli, Ayush Jain, Dong Huang Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12923v2 Resolving real-world human-object interactions in images is a many-to-many challenge, in which disentangling fine-grained concurrent physical contact is particularly difficult. Existing semantic contact estimation methods are either limited to single-human settings or require object geometries (e.g., meshes) in addition to the input image. Current state-of-the-art leverages powerful VLM for category-level semantics but struggles with multi-human scenarios and scales poorly in inference. We introduce Pi-HOC, a single-pass, instance-aware framework for dense 3D semantic contact prediction of all human-object pairs. Pi-HOC detects instances, creates dedicated human-object (HO) tokens for each pair, and refines them using an InteractionFormer. A SAM-based decoder then predicts dense contact on SMPL human meshes for each human-object pair. On the MMHOI and DAMON datasets, Pi-HOC significantly improves accuracy and localization over state-of-the-art methods while achieving 20x higher throughput. We further demonstrate that predicted contacts improve SAM-3D image-to-mesh reconstruction via a test-time optimization algorithm and enable referential contact prediction from language queries without additional training. --- ### MetFuse: Figurative Fusion between Metonymy and Metaphor Authors: Saptarshi Ghosh, Tianyu Jiang Published: 2026-04-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.12919v2 Metonymy and metaphor often co-occur in natural language, yet computational work has studied them largely in isolation. We introduce a framework that transforms a literal sentence into three figurative variants: metonymic, metaphoric, and hybrid. Using this framework, we construct MetFuse, the first dedicated dataset of figurative fusion between metonymy and metaphor, containing 1,000 human-verified meaning-aligned quadruplets totaling 4,000 sentences. Extrinsic experiments on eight existing benchmarks show that augmenting training data with MetFuse consistently improves both metonymy and metaphor classification, with hybrid examples yielding the largest gains on metonymy tasks. Using this dataset, we also analyze how the presence of one figurative type impacts another. Our findings show that both human annotators and large language models better identify metonymy in hybrid sentences than in metonymy-only sentences, demonstrating that the presence of a metaphor makes a metonymic noun more explicit. Our dataset is publicly available at: https://github.com/cincynlp/MetFuse. --- ### Radar-Camera BEV Multi-Task Learning with Cross-Task Attention Bridge for Joint 3D Detection and Segmentation Authors: Ahmet İnanç, Özgür Erkent Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12918v2 Bird's-eye-view (BEV) representations are the dominant paradigm for 3D perception in autonomous driving, providing a unified spatial canvas where detection and segmentation features are geometrically registered to the same physical coordinate system. However, existing radar-camera fusion methods treat these tasks in isolation, missing the opportunity for cross-task feature sharing: object-level geometric cues from detection can sharpen segmentation, while dense road-layout context from segmentation can anchor detection. We propose \textbf{CTAB} (Cross-Task Attention Bridge), a bidirectional module that exchanges features between detection and segmentation branches via multi-scale deformable attention in shared BEV space. CTAB is integrated into a multi-task framework with an Instance Normalization-based segmentation decoder and learnable BEV upsampling to provide a more detailed BEV representation. On nuScenes, CTAB improves segmentation on 7 classes over the joint multi-task baseline at essentially neutral detection. On a 4-class subset (drivable area, pedestrian crossing, walkway, vehicle), our joint multi-task model achieves 51.0 mIoU-4 while simultaneously providing competitive 3D detection. --- ### M3D-Stereo: A Multiple-Medium and Multiple-Degradation Dataset for Stereo Image Restoration Authors: Deqing Yang, Yingying Liu, Qicong Wang, Zhi Zeng, Dajiang Lu, Yibin Tian Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12917v1 Image restoration under adverse conditions, such as underwater, haze or fog, and low-light environments, remains a highly challenging problem due to complex physical degradations and severe information loss. Existing datasets are predominantly limited to a single degradation type or heavily rely on synthetic data without stereo consistency, inherently restricting their applicability in real-world scenarios. To address this, we introduce M3D-Stereo, a stereo dataset with 7904 high-resolution image pairs for image restoration research acquired in multiple media with multiple controlled degradation levels. It encompasses four degradation scenarios: underwater scatter, haze/fog, underwater low-light, and haze low-light. Each scenario forms a subset, and is divided into six levels of progressive degradation, allowing fine-grained evaluations of restoration methods with increasing severity of degradation. Collected via a laboratory setup, the dataset provides aligned stereo image pairs along with their pixel-wise consistent clear ground truths. Two restoration tasks, single-level and mixed-level degradation, were performed to verify its validity. M3D-Stereo establishes a better controlled and more realistic benchmark to evaluate image restoration and stereo matching methods in complex degradation environments. It is made public under LGPLv3 license. --- ### CoDe-R: Refining Decompiler Output with LLMs via Rationale Guidance and Adaptive Inference Authors: Qiang Zhang, Zhongnian Li Published: 2026-04-14 Categories: cs.SE, cs.AI, cs.CR arXiv: https://arxiv.org/abs/2604.12913v2 Binary decompilation is a critical reverse engineering task aimed at reconstructing high-level source code from stripped executables. Although Large Language Models (LLMs) have recently shown promise, they often suffer from "logical hallucinations" and "semantic misalignment" due to the irreversible semantic loss during compilation, resulting in generated code that fails to re-execute. In this study, we propose Cognitive Decompiler Refinement with Robustness (CoDe-R), a lightweight two-stage code refinement framework. The first stage introduces Semantic Cognitive Enhancement (SCE), a Rationale-Guided Semantic Injection strategy that trains the model to recover high-level algorithmic intent alongside code. The second stage introduces a Dynamic Dual-Path Fallback (DDPF) mechanism during inference, which adaptively balances semantic recovery and syntactic stability via a hybrid verification strategy. Evaluation on the HumanEval-Decompile benchmark demonstrates that CoDe-R (using a 1.3B backbone) establishes a new State-of-the-Art (SOTA) in the lightweight regime. Notably, it is the first 1.3B model to exceed an Average Re-executability Rate of 50.00%, significantly outperforming the baseline and effectively bridging the gap between efficient models and expert-level performance. Our code is available at https://github.com/Theaoi/CoDe-R. --- ### Round-Trip Translation Reveals What Frontier Multilingual Benchmarks Miss Authors: Ronald Skorobogat, Ameya Prabhu, Matthias Bethge Published: 2026-04-14 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.12911v1 Multilingual benchmarks guide the development of frontier models. Yet multilingual evaluations reported by frontier models are structured similar to popular reasoning and knowledge benchmarks, but across many languages. We show such benchmarks, and consequently multilingual evaluations, measure mathematical reasoning and factual recall, not multilingual proficiency. For example, thinking variants dramatically outperform instruct variants on these benchmarks, yet often perform worse on real-world multilingual tasks, such as LMArena. We propose a simple alternative: evaluate multilingual capability via round-trip translation. Given text in a source language, translate it to a target language and back; semantic gaps between the original and result expose failures in multilingual generation capabilities. Round-trip translation correlates almost perfectly (\r{ho} = 0.94) with user ratings on LMArena with our benchmark, requires no human reference translations, and does not require a more capable multilingual judge than tested models. Lastly, we introduce Lost in Translation (LiT), a challenging round-trip translation benchmark spanning widely spoken languages worldwide, for realistic evaluation of multilingual frontier models. --- ### Frequency-aware Decomposition Learning for Sensorless Wrench Forecasting on a Vibration-rich Hydraulic Manipulator Authors: Hyeonbeen Lee, Min-Jae Jung, Tae-Kyeong Yeu, Jong-Boo Han, Daegil Park, Jin-Gyun Kim Published: 2026-04-14 Categories: cs.RO, cs.LG arXiv: https://arxiv.org/abs/2604.12905v1 Force and torque (F/T) sensing is critical for robot-environment interaction, but physical F/T sensors impose constraints in size, cost, and fragility. To mitigate this, recent studies have estimated force/wrench sensorlessly from robot internal states. While existing methods generally target relatively slow interactions, tasks involving rapid interactions, such as grinding, can induce task-critical high-frequency vibrations, and estimation in such robotic settings remains underexplored. To address this gap, we propose a Frequency-aware Decomposition Network (FDN) for short-term forecasting of vibration-rich wrench from proprioceptive history. FDN predicts spectrally decomposed wrench with asymmetric deterministic and probabilistic heads, modeling the high-frequency residual as a learned conditional distribution. It further incorporates frequency-awareness to adaptively enhance input spectra with learned filtering and impose a frequency-band prior on the outputs. We pretrain FDN on a large-scale open-source robot dataset and transfer the learned proprioception-to-wrench representation to the downstream. On real-world grinding excavation data from a 6-DoF hydraulic manipulator and under a delayed estimation setting, FDN outperforms baseline estimators and forecasters in the high-frequency band and remains competitive in the low-frequency band. Transfer learning provides additional gains, suggesting the potential of large-scale pretraining and transfer learning for robotic wrench estimation. Code and data will be made available upon acceptance. --- ### A Sanity Check on Composed Image Retrieval Authors: Yikun Liu, Jiangchao Yao, Weidi Xie, Yanfeng Wang Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12904v1 Composed Image Retrieval (CIR) aims to retrieve a target image based on a query composed of a reference image, and a relative caption that specifies the desired modification. Despite the rapid development of CIR models, their performance is not well characterized by existing benchmarks, which inherently contain indeterminate queries degrading the evaluation (i.e., multiple candidate images, rather than solely the target image, meet the query criteria), and have not considered their effectiveness in the context of the multi-round system. Motivated by this, we consider improving the evaluation procedure from two aspects: 1) we introduce FISD, a Fully-Informed Semantically-Diverse benchmark, which employs generative models to precisely control the variables of reference-target image pairs, enabling a more accurate evaluation of CIR methods across six dimensions, without query ambiguity; 2) we propose an automatic multi-round agentic evaluation framework to probe the potential of the existing models in the interactive scenarios. By observing how models adapt and refine their choices over successive rounds of queries, this framework provides a more realistic appraisal of their efficacy in practical applications. Extensive experiments and comparisons prove the value of our novel evaluation on typical CIR methods. --- ### BEAM: Bi-level Memory-adaptive Algorithmic Evolution for LLM-Powered Heuristic Design Authors: Chuyang Xiang, Yichen Wei, Jiale Ma, Handing Wang, Junchi Yan Published: 2026-04-14 Categories: cs.AI, math.CO arXiv: https://arxiv.org/abs/2604.12898v1 Large Language Model-based Hyper Heuristic (LHH) has recently emerged as an efficient way for automatic heuristic design. However, most existing LHHs just perform well in optimizing a single function within a pre-defined solver. Their single-layer evolution makes them not effective enough to write a competent complete solver. While some variants incorporate hyperparameter tuning or attempt to generate complex code through iterative local modifications, they still lack a high-level algorithmic modeling, leading to limited exploration efficiency. To address this, we reformulate heuristic design as a Bi-level Optimization problem and propose \textbf{BEAM} (Bi-level Memory-adaptive Algorithmic Evolution). BEAM's exterior layer evolves high-level algorithmic structures with function placeholders through genetic algorithm (GA), while the interior layer realizes these placeholders via Monte Carlo Tree Search (MCTS). We further introduce an Adaptive Memory module to facilitate complex code generation. To support the evaluation for complex code generation, we point out the limitations of starting LHHs from scratch or from code templates and introduce a Knowledge Augmentation (KA) Pipeline. Experimental results on several optimization problems demonstrate that BEAM significantly outperforms existing LHHs, notably reducing the optimality gap by 37.84\% on aggregate in CVRP hybrid algorithm design. BEAM also designs a heuristic that outperforms SOTA Maximum Independent Set (MIS) solver KaMIS. --- ### Don't Show Pixels, Show Cues: Unlocking Visual Tool Reasoning in Language Models via Perception Programs Authors: Muhammad Kamran Janjua, Hugo Silva, Di Niu, Bahador Rashidi Published: 2026-04-14 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2604.12896v1 Multimodal language models (MLLMs) are increasingly paired with vision tools (e.g., depth, flow, correspondence) to enhance visual reasoning. However, despite access to these tool-generated visual cues, MLLMs often fail to benefit from them. Existing approaches typically feed raw tool outputs into the model, but these dense, pixel-level representations are misaligned with the language-native reasoning strengths of LLMs, leading to weak perception and reliance on language priors. We argue that, in problems where vision tools can provide the necessary visual cues, the bottleneck is not more tool calls or larger MLLMs, it is how tool outputs are represented. We introduce Perception Programs (P$^2$), a training-free, model-agnostic method that rewrites tool outputs into compact, structured, language-native summaries that MLLMs can directly parse and reason over. Across six perception-centric tasks in BLINK, P$^2$ consistently yields large improvements over base models and raw tool-augmented baselines. With GPT-5 Mini as the base model, P$^2$ raises its accuracy from 41.35\% to 86.47\% on multi-view reasoning, from 52.42\% to 81.45\% on relative depth, and achieves a 22\% average gain across tasks, setting new state-of-the-art results. Even on smaller MLLMs, e.g., InternVL3.5-4B and Qwen3VL-4B, we observe 15-40\% absolute gains from P$^2$, surpassing prior agentic, supervised, and RL-based tool-use methods-without any training or model modifications. --- ### Representing 3D Faces with Learnable B-Spline Volumes Authors: Prashanth Chandran, Daoye Wang, Timo Bolkart Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12894v1 We present CUBE (Control-based Unified B-spline Encoding), a new geometric representation for human faces that combines B-spline volumes with learned features, and demonstrate its use as a decoder for 3D scan registration and monocular 3D face reconstruction. Unlike existing B-spline representations with 3D control points, CUBE is parametrized by a lattice (e.g., 8 x 8 x 8) of high-dimensional control features, increasing the model's expressivity. These features define a continuous, two-stage mapping from a 3D parametric domain to 3D Euclidean space via an intermediate feature space. First, high-dimensional control features are locally blended using the B-spline bases, yielding a high-dimensional feature vector whose first three values define a 3D base mesh. A small MLP then processes this feature vector to predict a residual displacement from the base shape, yielding the final refined 3D coordinates. To reconstruct 3D surfaces in dense semantic correspondence, CUBE is queried at 3D coordinates sampled from a fixed template mesh. Crucially, CUBE retains the local support property of traditional B-spline representations, enabling local surface editing by updating individual control features. We demonstrate the strengths of this representation by training transformer-based encoders to predict CUBE's control features from unstructured point clouds and monocular images, achieving state-of-the-art scan registration results compared to recent baselines. --- ### TCL: Enabling Fast and Efficient Cross-Hardware Tensor Program Optimization via Continual Learning Authors: Chaoyao Shen, Linfeng Jiang, Yixian Shen, Tao Xu, Guoqing Li, Anuj Pathania, Andy D. Pimentel, Meng Zhang Published: 2026-04-14 Categories: cs.LG, cs.AR arXiv: https://arxiv.org/abs/2604.12891v1 Deep learning (DL) compilers rely on cost models and auto-tuning to optimize tensor programs for target hardware. However, existing approaches depend on large offline datasets, incurring high collection costs and offering suboptimal transferability across platforms. In this paper, we introduce TCL, a novel efficient and transferable compiler framework for fast tensor program optimization across diverse hardware platforms to address these challenges. Specifically, TCL is built on three core enablers: (1) the RDU Sampler, a data-efficient active learning strategy that selects only 10% of tensor programs by jointly optimizing Representativeness, Diversity, and Uncertainty, substantially reducing data collection costs while maintaining near-original model accuracy; (2) a new Mamba-based cost model that efficiently captures long-range schedule dependencies while achieving a favorable trade-off between prediction accuracy and computational cost through reduced parameterization and lightweight sequence modeling; and (3) a continuous knowledge distillation framework that effectively and progressively transfers knowledge across multiple hardware platforms while avoiding the parameter explosion and data dependency issues typically caused by traditional multi-task learning. Extensive experiments validate the effectiveness of each individual enabler and the holistic TCL framework. When optimizing a range of mainstream DL models on both CPU and GPU platforms, TCL achieves, on average, 16.8x and 12.48x faster tuning time, and 1.20x and 1.13x lower inference latency, respectively, compared to Tenset-MLP. --- ### Towards Long-horizon Agentic Multimodal Search Authors: Yifan Du, Zikang Liu, Jinbiao Peng, Jie Wu, Junyi Li, Jinyang Li, Wayne Xin Zhao, Ji-Rong Wen Published: 2026-04-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.12890v2 Multimodal deep search agents have shown great potential in solving complex tasks by iteratively collecting textual and visual evidence. However, managing the heterogeneous information and high token costs associated with multimodal inputs over long horizons remains a critical challenge, as existing methods often suffer from context explosion or the loss of crucial visual signals. To address this, we propose a novel Long-horizon MultiModal deep search framework, named LMM-Searcher, centered on a file-based visual representation mechanism. By offloading visual assets to an external file system and mapping them to lightweight textual identifiers (UIDs), our approach mitigates context overhead while preserving multimodal information for future access. We equip the agent with a tailored fetch-image tool, enabling a progressive, on-demand visual loading strategy for active perception. Furthermore, we introduce a data synthesis pipeline designed to generate queries requiring complex cross-modal multi-hop reasoning. Using this pipeline, we distill 12K high-quality trajectories to fine-tune Qwen3-VL-Thinking-30A3B into a specialized multimodal deep search agent. Extensive experiments across four benchmarks demonstrate that our method successfully scales to 100-turn search horizons, achieving state-of-the-art performance among open-source models on challenging long-horizon benchmarks like MM-BrowseComp and MMSearch-Plus, while also exhibiting strong generalizability across different base models. Our code will be released in https://github.com/RUCAIBox/LMM-Searcher. --- ### VideoFlexTok: Flexible-Length Coarse-to-Fine Video Tokenization Authors: Andrei Atanov, Jesse Allardice, Roman Bachmann, Oğuzhan Fatih Kar, R Devon Hjelm, David Griffiths, Peter Fu, Afshin Dehghan, Amir Zamir Published: 2026-04-14 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2604.12887v1 Visual tokenizers map high-dimensional raw pixels into a compressed representation for downstream modeling. Beyond compression, tokenizers dictate what information is preserved and how it is organized. A de facto standard approach to video tokenization is to represent a video as a spatiotemporal 3D grid of tokens, each capturing the corresponding local information in the original signal. This requires the downstream model that consumes the tokens, e.g., a text-to-video model, to learn to predict all low-level details "pixel-by-pixel" irrespective of the video's inherent complexity, leading to high learning complexity. We present VideoFlexTok, which represents videos with a variable-length sequence of tokens structured in a coarse-to-fine manner -- where the first tokens (emergently) capture abstract information, such as semantics and motion, and later tokens add fine-grained details. The generative flow decoder enables realistic video reconstructions from any token count. This representation structure allows adapting the token count according to downstream needs and encoding videos longer than the baselines with the same budget. We evaluate VideoFlexTok on class- and text-to-video generative tasks and show that it leads to more efficient training compared to 3D grid tokens, e.g., achieving comparable generation quality (gFVD and ViCLIP Score) with a 5x smaller model (1.1B vs 5.2B). Finally, we demonstrate how VideoFlexTok can enable long video generation without prohibitive computational cost by training a text-to-video model on 10-second 81-frame videos with only 672 tokens, 8x fewer than a comparable 3D grid tokenizer. --- ### An abstract model of nonrandom, non-Lamarckian mutation in evolution using a multivariate estimation-of-distribution algorithm Authors: Liudmyla Vasylenko, Adi Livnat Published: 2026-04-14 Categories: cs.NE, q-bio.PE arXiv: https://arxiv.org/abs/2604.12884v1 At the fundamental conceptual level, two alternatives have traditionally been considered for how mutations arise and how evolution happens: 1) random mutation and natural selection, and 2) Lamarckism. Recently, the theory of Interaction-based Evolution (IBE) has been proposed, according to which mutations are neither random nor Lamarckian, but are influenced by information accumulating internally in the genome over generations. Based on the estimation-of-distribution algorithms framework, we present a simulation model that demonstrates nonrandom, non-Lamarckian mutation concretely while capturing indirectly several aspects of IBE: selection, recombination, and nonrandom, non-Lamarckian mutation interact in a complementary fashion; evolution is driven by the interaction of parsimony and fit; and random bits do not directly encode improvement but enable generalization by the manner in which they connect with the rest of the evolutionary process. Connections are drawn to Darwin's observations that changed conditions increase the rate of production of heritable variation; to the causes of bell-shaped distributions of traits and how these distributions respond to selection; and to computational learning theory, where analogizing evolution to learning in accord with IBE casts individuals as examples and places the learned hypothesis at the population level. The model highlights the importance of incorporating internal integration of information through heritable change in both evolutionary theory and evolutionary computation. --- ### FastGrasp: Learning-based Whole-Body Control Method for Fast Dexterous Grasping with Mobile Manipulators Authors: Heng Tao, Yiming Zhong, Zemin Yang, Yuexin Ma Published: 2026-04-14 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2604.12879v2 Fast dexterous grasping while a mobile base remains in motion requires coordinated whole-body control and rapid adaptation to physical contact. We propose FastGrasp, a two-stage learning framework that integrates grasp guidance, whole-body control, and tactile feedback. First, a pretrained conditional variational autoencoder generates diverse grasp candidates from object point clouds. Candidates are filtered by approach direction and supporting-surface constraints, then ranked using an envelopment-based criterion combining grasp width and depth coverage measures. Second, a reinforcement learning policy jointly controls the mobile base, arm, and dexterous hand, using the selected grasp as guidance and tactile observations for online grasp adjustment. The policy is trained with domain randomization and deployed with command filtering and tactile-triggered grasp tightening. Simulation experiments show higher grasp success rates than the evaluated baselines under full and partial point-cloud observations, while real-world experiments demonstrate sim-to-real transfer across diverse object geometries. --- ### FastGrasp: Learning-based Whole-body Control method for Fast Dexterous Grasping with Mobile Manipulators Authors: Heng Tao, Yiming Zhong, Zemin Yang, Yuexin Ma Published: 2026-04-14 Categories: cs.RO, cs.AI arXiv: https://arxiv.org/abs/2604.12879v1 Fast grasping is critical for mobile robots in logistics, manufacturing, and service applications. Existing methods face fundamental challenges in impact stabilization under high-speed motion, real-time whole-body coordination, and generalization across diverse objects and scenarios, limited by fixed bases, simple grippers, or slow tactile response capabilities. We propose \textbf{FastGrasp}, a learning-based framework that integrates grasp guidance, whole-body control, and tactile feedback for mobile fast grasping. Our two-stage reinforcement learning strategy first generates diverse grasp candidates via conditional variational autoencoder conditioned on object point clouds, then executes coordinated movements of mobile base, arm, and hand guided by optimal grasp selection. Tactile sensing enables real-time grasp adjustments to handle impact effects and object variations. Extensive experiments demonstrate superior grasping performance in both simulation and real-world scenarios, achieving robust manipulation across diverse object geometries through effective sim-to-real transfer. --- ### AISafetyBenchExplorer: A Metric-Aware Catalogue of AI Safety Benchmarks Reveals Fragmented Measurement and Weak Benchmark Governance Authors: Abiodun A. Solanke Published: 2026-04-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.12875v2 The rapid expansion of large language model (LLM) safety evaluation has produced a substantial benchmark ecosystem, but not a correspondingly coherent measurement ecosystem. We present AISafetyBenchExplorer, a structured catalogue of 195 AI safety benchmarks released between 2018 and 2026, organized through a multi-sheet schema that records benchmark-level metadata, metric-level definitions, benchmark-paper metadata, and repository activity. This design enables meta-analysis not only of what benchmarks exist, but also of how safety is operationalized, aggregated, and judged across the literature. Using the updated catalogue, we identify a central structural problem: benchmark proliferation has outpaced measurement standardization. The current landscape is dominated by medium-complexity benchmarks (94/195), while only 7 benchmarks occupy the Popular tier. The workbook further reports strong concentration around English-only evaluation (165/195), evaluation-only resources (170/195), stale GitHub repositories (137/195), stale Hugging Face datasets (96/195), and heavy reliance on arXiv preprints among benchmarks with known venue metadata. At the metric level, the catalogue shows that familiar labels such as accuracy, F1 score, safety score, and aggregate benchmark scores often conceal materially different judges, aggregation rules, and threat models. We argue that the field's main failure mode is fragmentation rather than scarcity. Researchers now have many benchmark artifacts, but they often lack a shared measurement language, a principled basis for benchmark selection, and durable stewardship norms for post publication maintenance. AISafetyBenchExplorer addresses this gap by providing a traceable benchmark catalogue, a controlled metadata schema, and a complexity taxonomy that together support more rigorous benchmark discovery, comparison, and meta-evaluation. --- ### LIFE -- an energy efficient advanced continual learning agentic AI framework for frontier systems Authors: Anne Lee, Gurudutt Hosangadi Published: 2026-04-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.12874v1 The rapid advancement of AI has changed the character of HPC usage such as dimensioning, provisioning, and execution. Not only has energy demand been amplified, but existing rudimentary continual learning capabilities limit ability of AI to effectively manage HPCs. This paper reviews emerging directions beyond monolithic transformers, emphasizing agentic AI and brain inspired architectures as complementary paths toward sustainable, adaptive systems. We propose LIFE, a reasoning and Learning framework that is Incremental, Flexible, and Energy efficient that is implemented as an agent centric system rather than a single monolithic model. LIFE uniquely combines four components to realize self evolving network management and operations in HPCs. The components are an orchestrator, Agentic Context Engineering, a novel memory system, and information lattice learning. LIFE can also generalize to enable a variety of orthogonal use cases. We ground LIFE in a specific closed loop HPC operations example for detecting and mitigating latency spikes experienced by critical micro services running on a Kubernetes like cluster. --- ### QuarkMedSearch: A Long-Horizon Deep Search Agent for Exploring Medical Intelligence Authors: Zhichao Lin, Zhichao Liang, Gaoqiang Liu, Meng Xu, Baoyu Xiang, Shuxin Zhao, Yao Wu, Jian Xu, Guanjun Jiang Published: 2026-04-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.12867v4 As agentic foundation models continue to evolve, how to further improve their performance in vertical domains has become an important challenge. To this end, building upon Tongyi DeepResearch, a powerful agentic foundation model, we focus on the Chinese medical deep search scenario and propose QuarkMedSearch, systematically exploring a full-pipeline approach spanning medical multi-hop data construction, training strategies, and evaluation benchmarks to further push and assess its performance upper bound in vertical domains. Specifically, for data synthesis, to address the scarcity of deep search training data in the medical domain, we combine a large-scale medical knowledge graph with real-time online exploration to construct long-horizon medical deep search training data; for post-training, we adopt a two-stage SFT and RL training strategy that progressively enhances the model's planning, tool invocation, and reflection capabilities required for deep search, while maintaining search efficiency; for evaluation, we collaborate with medical experts to construct the QuarkMedSearch Benchmark through rigorous manual verification. Experimental results demonstrate that QuarkMedSearch achieves state-of-the-art performance among open-source models of comparable scale on the QuarkMedSearch Benchmark, while also maintaining strong competitiveness on general benchmarks. --- ### From edges to meaning: Semantic line sketches as a cognitive scaffold for ancient pictograph invention Authors: Seowung Leem, Lin Gu, Ruogu Fang Published: 2026-04-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.12865v2 Humans readily recognize objects from sparse line drawings, a capacity that appears early in development and persists across cultures, suggesting neural rather than purely learned origins. Yet the computational mechanism by which the brain transforms high-level semantic knowledge into low-level visual symbols remains poorly understood. Here we propose that ancient pictographic writing emerged from the brain's intrinsic tendency to compress visual input into stable, boundary-based abstractions. We construct a biologically inspired digital twin of the visual hierarchy that encodes an image into low-level features, generates a contour sketch, and iteratively refines it through top-down feedback guided by semantic representations, mirroring the feedforward and recurrent architecture of the human visual cortex. The resulting symbols bear striking structural resemblance to early pictographs across culturally distant writing systems, including Egyptian hieroglyphs, Chinese oracle bone characters, and proto-cuneiform, and offer candidate interpretations for undeciphered scripts. Our findings support a neuro-computational origin of pictographic writing and establish a framework in which AI can recapitulate the cognitive processes by which humans first externalized perception into symbols. --- ### Artificial Intelligence for Modeling and Simulation of Mixed Automated and Human Traffic Authors: Saeed Rahmani, Shiva Rasouli, Daphne Cornelisse, Eugene Vinitsky, Bart van Arem, Simeon C. Calvert Published: 2026-04-14 Categories: cs.AI, cs.RO, eess.SY arXiv: https://arxiv.org/abs/2604.12857v1 Autonomous vehicles (AVs) are now operating on public roads, which makes their testing and validation more critical than ever. Simulation offers a safe and controlled environment for evaluating AV performance in varied conditions. However, existing simulation tools mainly focus on graphical realism and rely on simple rule-based models and therefore fail to accurately represent the complexity of driving behaviors and interactions. Artificial intelligence (AI) has shown strong potential to address these limitations; however, despite the rapid progress across AI methodologies, a comprehensive survey of their application to mixed autonomy traffic simulation remains lacking. Existing surveys either focus on simulation tools without examining the AI methods behind them, or cover ego-centric decision-making without addressing the broader challenge of modeling surrounding traffic. Moreover, they do not offer a unified taxonomy of AI methods covering individual behavior modeling to full scene simulation. To address these gaps, this survey provides a structured review and synthesis of AI methods for modeling AV and human driving behavior in mixed autonomy traffic simulation. We introduce a taxonomy that organizes methods into three families: agent-level behavior models, environment-level simulation methods, and cognitive and physics-informed methods. The survey analyzes how existing simulation platforms fall short of the needs of mixed autonomy research and outlines directions to narrow this gap. It also provides a chronological overview of AI methods and reviews evaluation protocols and metrics, simulation tools, and datasets. By covering both traffic engineering and computer science perspectives, we aim to bridge the gap between these two communities. --- ### PianoFlow: Music-Aware Streaming Piano Motion Generation with Bimanual Coordination Authors: Xuan Wang, Kai Ruan, Jiayi Han, Kaiyue Zhou, Gaoang Wang Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12856v2 Audio-driven bimanual piano motion generation requires precise modeling of complex musical structures and dynamic cross-hand coordination. However, existing methods often rely on acoustic-only representations lacking symbolic priors, employ inflexible interaction mechanisms, and are limited to computationally expensive short-sequence generation. To address these limitations, we propose PianoFlow, a flow-matching framework for precise and coordinated bimanual piano motion synthesis. Our approach strategically leverages MIDI as a privileged modality during training, distilling these structured musical priors to achieve deep semantic understanding while maintaining audio-only inference. Furthermore, we introduce an asymmetric role-gated interaction module to explicitly capture dynamic cross-hand coordination through role-aware attention and temporal gating. To enable real-time streaming generation for arbitrarily long sequences, we design an autoregressive flow continuation scheme that ensures seamless cross-chunk temporal coherence. Extensive experiments on the PianoMotion10M dataset demonstrate that PianoFlow achieves superior quantitative and qualitative performance, while accelerating inference by over 9\times compared to previous methods. --- ### Growing Pains: Extensible and Efficient LLM Benchmarking Via Fixed Parameter Calibration Authors: Eliya Habba, Itay Itzhak, Asaf Yehudai, Yotam Perlitz, Elron Bandel, Michal Shmueli-Scheuer, Leshem Choshen, Gabriel Stanovsky Published: 2026-04-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.12843v3 The rapid release of both language models and benchmarks makes it increasingly costly to evaluate every model on every dataset. In practice, models are often evaluated on different samples, making scores difficult to compare across studies. To address this, we propose a framework based on multidimensional Item Response Theory (IRT) that uses anchor items to calibrate new benchmarks to the evaluation suite while holding previously calibrated item parameters fixed. Our approach supports a realistic evaluation setting in which datasets are introduced over time and models are evaluated only on the datasets available at the time of evaluation, while a fixed anchor set for each dataset is used so that results from different evaluation periods can be compared directly. In large-scale experiments on more than 400 models, our framework predicts full-evaluation performance within 2-3 percentage points using only 100 anchor questions per dataset, with Spearman $ρ\geq 0.9$ for ranking preservation preservation. This shows that benchmark suites can grow over time while preserving score comparability, since adding a new dataset requires running existing models only on that dataset's anchors. Code and data are available at: https://eliyahabba.github.io/growing-pains/ --- ### Growing Pains: Extensible and Efficient LLM Benchmarking Via Fixed Parameter Calibration Authors: Eliya Habba, Itay Itzhak, Asaf Yehudai, Yotam Perlitz, Elron Bandel, Michal Shmueli-Scheuer, Leshem Choshen, Gabriel Stanovsky Published: 2026-04-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.12843v2 The rapid release of both language models and benchmarks makes it increasingly costly to evaluate every model on every dataset. In practice, models are often evaluated on different samples, making scores difficult to compare across studies. To address this, we propose a framework based on multidimensional Item Response Theory (IRT) that uses anchor items to calibrate new benchmarks to the evaluation suite while holding previously calibrated item parameters fixed. Our approach supports a realistic evaluation setting in which datasets are introduced over time and models are evaluated only on the datasets available at the time of evaluation, while a fixed anchor set for each dataset is used so that results from different evaluation periods can be compared directly. In large-scale experiments on more than $400$ models, our framework predicts full-evaluation performance within 2-3 percentage points using only $100$ anchor questions per dataset, with Spearman $ρ\geq 0.9$ for ranking preservation, showing that it is possible to extend benchmark suites over time while preserving score comparability, at a constant evaluation cost per new dataset. Code available at https://github.com/eliyahabba/growing-pains --- ### Rapid LoRA Aggregation for Wireless Channel Adaptation in Open-Set Radio Frequency Fingerprinting Authors: Mingxi Zhang, Renjie Xie, Jincheng Wang, Guyue Li, Wei Xu Published: 2026-04-14 Categories: eess.SP, cs.CR, cs.LG arXiv: https://arxiv.org/abs/2604.12834v1 Radio frequency fingerprints (RFFs) enable secure wireless authentication but struggle in open-set scenarios with unknown devices and varying channels. Existing methods face challenges in generalization and incur high computational costs. We propose a lightweight, self-adaptive RFF extraction framework using Low-Rank Adaptation (LoRA). By pretraining LoRA modules per environment, our method enables fast adaptation to unseen channel conditions without full retraining. During inference, a weighted combination of LoRAs dynamically enhances feature extraction. Experimental results demonstrate a 15% reduction in equal error rate (EER) compared to non-finetuned baselines and an 83% decrease in training time relative to full fine-tuning, using the same training dataset. This approach provides a scalable and efficient solution for open-set RFF authentication in dynamic wireless vehicular networks. --- ### Challenging Vision-Language Models with Physically Deployable Multimodal Semantic Lighting Attacks Authors: Yingying Zhao, Chengyin Hu, Qike Zhang, Xin Li, Xin Wang, Yiwei Wei, Jiujiang Guo, Jiahuan Long, Tingsong Jiang, Wen Yao Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12833v1 Vision-Language Models (VLMs) have shown remarkable performance, yet their security remains insufficiently understood. Existing adversarial studies focus almost exclusively on the digital setting, leaving physical-world threats largely unexplored. As VLMs are increasingly deployed in real environments, this gap becomes critical, since adversarial perturbations must be physically realizable. Despite this practical relevance, physical attacks against VLMs have not been systematically studied. Such attacks may induce recognition failures and further disrupt multimodal reasoning, leading to severe semantic misinterpretation in downstream tasks. Therefore, investigating physical attacks on VLMs is essential for assessing their real-world security risks. To address this gap, we propose Multimodal Semantic Lighting Attacks (MSLA), the first physically deployable adversarial attack framework against VLMs. MSLA uses controllable adversarial lighting to disrupt multimodal semantic understanding in real scenes, attacking semantic alignment rather than only task-specific outputs. Consequently, it degrades zero-shot classification performance of mainstream CLIP variants while inducing severe semantic hallucinations in advanced VLMs such as LLaVA and BLIP across image captioning and visual question answering (VQA). Extensive experiments in both digital and physical domains demonstrate that MSLA is effective, transferable, and practically realizable. Our findings provide the first evidence that VLMs are highly vulnerable to physically deployable semantic attacks, exposing a previously overlooked robustness gap and underscoring the urgent need for physical-world robustness evaluation of VLMs. --- ### Detecting and refurbishing ground truth errors during training of deep learning-based echocardiography segmentation models Authors: Iman Islam, Bram Ruijsink, Andrew J. Reader, Andrew P. King Published: 2026-04-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.12832v1 Deep learning-based medical image segmentation typically relies on ground truth (GT) labels obtained through manual annotation, but these can be prone to random errors or systematic biases. This study examines the robustness of deep learning models to such errors in echocardiography (echo) segmentation and evaluates a novel strategy for detecting and refurbishing erroneous labels during model training. Using the CAMUS dataset, we simulate three error types, then compare a loss-based GT label error detection method with one based on Variance of Gradients (VOG). We also propose a pseudo-labelling approach to refurbish suspected erroneous GT labels. We assess the performance of our proposed approach under varying error levels. Results show that VOG proved highly effective in flagging erroneous GT labels during training. However, a standard U-Net maintained strong performance under random label errors and moderate levels of systematic errors (up to 50%). The detection and refurbishment approach improved performance, particularly under high-error conditions. --- ### Loop Corrections in Random Feature Models: Training Error and Generalization Gap Authors: Taeyoung Kim Published: 2026-04-14 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2604.12827v4 We study fixed-design random feature ridge regression beyond the mean-kernel approximation. The expectation is taken over the frozen-feature ensemble, conditional on the training sample. Because the predictor is a nonlinear function of the empirical kernel, its mean training error, test error, and conditional generalization gap depend on centered kernel covariances as well as on mean kernel objects. We derive the covariance-level, or one-loop, corrections from a finite resolvent identity. This avoids an almost-sure Neumann-series assumption and yields an explicit remainder bound for the training error. The test correction additionally requires mixed train--test covariance tensors and therefore cannot, in general, be recovered from the train-restricted kernel law alone. Numerical experiments show that the covariance term removes the leading inverse-width discrepancy in a controlled regularization regime, while an ablation that discards the mixed train--test covariance fails for the test error even though the training correction is unchanged. They also identify a width--regularization boundary, measured by a resolvent-fluctuation diagnostic, beyond which the second-order truncation loses accuracy. All inverse-width statements are for fixed sample size, depth, design, and positive kernel-level regularization. The analysis concerns frozen features and does not model feature learning or neural-tangent-kernel evolution. --- ### Loop Corrections to the Training Error and Generalization Gap of Random Feature Models Authors: Taeyoung Kim Published: 2026-04-14 Categories: cs.LG, cs.AI, stat.ML arXiv: https://arxiv.org/abs/2604.12827v3 We investigate random feature models in which neural networks sampled from a prescribed initialization ensemble are frozen and used as random features, with only the readout weights optimized. Adopting a statistical-physics viewpoint, we study the training error, test error, and generalization gap beyond the mean kernel approximation. Since the predictor is a nonlinear functional of the induced random kernel, the ensemble-averaged errors depend not only on the mean kernel but also on higher-order fluctuation statistics. Within an effective field-theoretic framework, these finite-width contributions naturally appear as loop corrections. We derive loop corrections to the training error, test error, and generalization gap, obtain their scaling laws, and support the theory with experimental verification. --- ### RePAIR: Interactive Machine Unlearning through Prompt-Aware Model Repair Authors: Jagadeesh Rachapudi, Pranav Singh, Ritali Vatsi, Praful Hambarde, Amit Shukla Published: 2026-04-14 Categories: cs.AI, cs.CL arXiv: https://arxiv.org/abs/2604.12820v1 Large language models (LLMs) inherently absorb harmful knowledge, misinformation, and personal data during pretraining on large-scale web corpora, with no native mechanism for selective removal. While machine unlearning offers a principled solution, existing approaches are provider-centric, requiring retraining pipelines, curated retain datasets, and direct intervention by model service providers (MSPs), thereby excluding end users from controlling their own data. We introduce Interactive Machine Unlearning (IMU), a new paradigm in which users can instruct LLMs to forget targeted knowledge through natural language at inference time. To realize IMU, we propose RePAIR, a prompt-aware model repair framework comprising (i) a watchdog model for unlearning intent detection, (ii) a surgeon model for generating repair procedures, and (iii) a patient model whose parameters are updated autonomously. At the core of RePAIR, we develop Steering Through Activation Manipulation with PseudoInverse (STAMP), a training-free, single-sample unlearning method that redirects MLP activations toward a refusal subspace via closed-form pseudoinverse updates. Its low-rank variant reduces computational complexity from O(d^3) to O(r^3 + r^2 * d), enabling efficient on-device unlearning with up to ~3x speedup over training-based baselines. Extensive experiments across harmful knowledge suppression, misinformation correction, and personal data erasure demonstrate that RePAIR achieves near-zero forget scores (Acc_f = 0.00, F-RL = 0.00) while preserving model utility (Acc_r up to 84.47, R-RL up to 0.88), outperforming six state-of-the-art baselines. These results establish RePAIR as an effective and practical framework for user-driven model editing, advancing transparent and on-device control over learned knowledge, with potential extensions to multimodal foundation models. --- ### Understanding and Improving Continuous Adversarial Training for LLMs via In-context Learning Theory Authors: Shaopeng Fu, Di Wang Published: 2026-04-14 Categories: cs.LG, cs.CR, stat.ML arXiv: https://arxiv.org/abs/2604.12817v1 Adversarial training (AT) is an effective defense for large language models (LLMs) against jailbreak attacks, but performing AT on LLMs is costly. To improve the efficiency of AT for LLMs, recent studies propose continuous AT (CAT) that searches for adversarial inputs within the continuous embedding space of LLMs during AT. While CAT has achieved empirical success, its underlying mechanism, i.e., why adversarial perturbations in the embedding space can help LLMs defend against jailbreak prompts synthesized in the input token space, remains unknown. This paper presents the first theoretical analysis of CAT on LLMs based on in-context learning (ICL) theory. For linear transformers trained with adversarial examples from the embedding space on in-context linear regression tasks, we prove a robust generalization bound that has a negative correlation with the perturbation radius in the embedding space. This clearly explains why CAT can defend against jailbreak prompts from the LLM's token space. Further, the robust bound shows that the robustness of an adversarially trained LLM is closely related to the singular values of its embedding matrix. Based on this, we propose to improve LLM CAT by introducing an additional regularization term, which depends on singular values of the LLM's embedding matrix, into the objective function of CAT. Experiments on real-world LLMs demonstrate that our method can help LLMs achieve a better jailbreak robustness-utility tradeoff. The code is available at https://github.com/fshp971/continuous-adv-icl. --- ### The role of System 1 and System 2 semantic memory structure in human and LLM biases Authors: Katherine Abramski, Giulio Rossetti, Massimo Stella Published: 2026-04-14 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.12816v1 Implicit biases in both humans and large language models (LLMs) pose significant societal risks. Dual process theories propose that biases arise primarily from associative System 1 thinking, while deliberative System 2 thinking mitigates bias, but the cognitive mechanisms that give rise to this phenomenon remain poorly understood. To better understand what underlies this duality in humans, and possibly in LLMs, we model System 1 and System 2 thinking as semantic memory networks with distinct structures, built from comparable datasets generated by both humans and LLMs. We then investigate how these distinct semantic memory structures relate to implicit gender bias using network-based evaluation metrics. We find that semantic memory structures are irreducible only in humans, suggesting that LLMs lack certain types of human-like conceptual knowledge. Moreover, semantic memory structure relates consistently to implicit bias only in humans, with lower levels of bias in System~2 structures. These findings suggest that certain types of conceptual knowledge contribute to bias regulation in humans, but not in LLMs, highlighting fundamental differences between human and machine cognition. --- ### DPC-VQA: Decoupling Quality Perception and Residual Calibration for Video Quality Assessment Authors: Xinyue Li, Shubo Xu, Zhichao Zhang, Zhaolin Cai, Yitong Chen, Guangtao Zhai Published: 2026-04-14 Categories: cs.CV, cs.MM arXiv: https://arxiv.org/abs/2604.12813v2 Recent multimodal large language models (MLLMs) have shown promising performance on video quality assessment (VQA) tasks. However, adapting them to new scenarios remains expensive due to large-scale retraining and costly mean opinion score (MOS) annotations. In this paper, we argue that a pretrained MLLM already provides a useful perceptual prior for VQA, and that the main challenge is to efficiently calibrate this prior to the target MOS space. Based on this insight, we propose DPC-VQA, a decoupling perception and calibration framework for video quality assessment. Specifically, DPC-VQA uses a frozen MLLM to provide a base quality estimate and perceptual prior, and employs a lightweight calibration branch to predict a residual correction for target-scenario adaptation. This design avoids costly end-to-end retraining while maintaining reliable performance with lower training and data costs. Extensive experiments on both user-generated content (UGC) and AI-generated content (AIGC) benchmarks show that DPC-VQA achieves competitive performance against representative baselines, while using less than 2% of the trainable parameters of conventional MLLM-based VQA methods and remaining effective with only 20% of MOS labels. The code will be released upon publication. --- ### DocSeeker: Structured Visual Reasoning with Evidence Grounding for Long Document Understanding Authors: Hao Yan, Yuliang Liu, Xingchen Liu, Yuyi Zhang, Minghui Liao, Jihao Wu, Wei Chen, Xiang Bai Published: 2026-04-14 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.12812v5 Existing Multimodal Large Language Models (MLLMs) suffer from significant performance degradation on the long document understanding task as document length increases. This stems from two fundamental challenges: 1) a low Signal-to-Noise Ratio (SNR), with crucial evidence buried in irrelevant pages; and 2) supervision scarcity, as datasets offering only final short answers provide a weak learning signal. In this paper, we address these challenges by proposing a paradigm that requires the model to execute a structured Analysis, Localization and Reasoning workflow. To instill this capability, we design a two-stage training framework: we first perform Supervised Fine-Tuning on high-quality data generated via an efficient knowledge distillation strategy. Subsequently, we employ an Evidence-aware Group Relative Policy Optimization which jointly optimizes for both evidence localization and answer accuracy. Additionally, we introduce a Evidence-Guided Resolution Allocation strategy to mitigate memory constraints of training on multi-pages documents. Extensive experiments demonstrate that DocSeeker achieves superior performance on both in-domain and out-of-domain tasks. We show it robustly generalizes from short-page training to ultra-long documents and is naturally synergistic with visual Retrieval-Augmented Generation systems, serving as a solid foundation for their implementation. --- ### Algorithmic Analysis of Dense Associative Memory: Finite-Size Guarantees and Adversarial Robustness Authors: Madhava Gaikwad Published: 2026-04-14 Categories: cs.LG, cs.AI, cs.NE arXiv: https://arxiv.org/abs/2604.12811v1 Dense Associative Memory (DAM) generalizes Hopfield networks through higher-order interactions and achieves storage capacity that scales as $O(N^{n-1})$ under suitable pattern separation conditions. Existing dynamical analyses primarily study the thermodynamic limit $N\to\infty$ with randomly sampled patterns and therefore do not provide finite-size guarantees or explicit convergence rates. We develop an algorithmic analysis of DAM retrieval dynamics that yields finite-$N$ guarantees under explicit, verifiable pattern conditions. Under a separation assumption and a bounded-interference condition at high loading, we prove geometric convergence of asynchronous retrieval dynamics, which implies $O(\log N)$ convergence time once the trajectory enters the basin of attraction. We further establish adversarial robustness bounds expressed through an explicit margin condition that quantifies the number of corrupted bits tolerable per sweep, and derive capacity guarantees that scale as $Θ(N^{n-1})$ up to polylogarithmic factors in the worst case, while recovering the classical $Θ(N^{n-1})$ scaling for random pattern ensembles. Finally, we show that DAM retrieval dynamics admit a potential-game interpretation that ensures convergence to pure Nash equilibria under asynchronous updates. Complete proofs are provided in the appendices, together with preliminary experiments that illustrate the predicted convergence, robustness, and capacity scaling behavior. --- ### Rethinking Satellite Image Restoration for Onboard AI: A Lightweight Learning-Based Approach Authors: Adrien Dorise, Marjorie Bellizzi, Omar Hlimi Published: 2026-04-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.12807v2 Satellite image restoration aims to improve image quality by compensating for degradations (e.g., noise and blur) introduced by the imaging system and acquisition conditions. As a fundamental preprocessing step, restoration directly impacts both ground-based product generation and emerging onboard AI applications. Traditional restoration pipelines based on sequential physical models are computationally intensive and slow, making them unsuitable for onboard environments. In this paper, we introduce ConvBEERS: a Convolutional Board-ready Embedded and Efficient Restoration model for Space to investigate whether a light and non-generative residual convolutional network, trained on simulated satellite data, can match or surpass a traditional ground-processing restoration pipeline across multiple operating conditions. Experiments conducted on simulated datasets and real Pleiades-HR imagery demonstrate that the proposed approach achieves competitive image quality, with a +6.9dB PSNR improvement. Evaluation on a downstream object detection task demonstrates that restoration significantly improves performance, with up to +5.1% mAP@50. In addition, successful deployment on a Xilinx Versal VCK190 FPGA validates its practical feasibility for satellite onboard processing, with a ~41x reduction in latency compared to the traditional pipeline. These results demonstrate the relevance of using lightweight CNNs to achieve competitive restoration quality while addressing real-world constraints in spaceborne systems. --- ### Interpretable Relational Inference with LLM-Guided Symbolic Dynamics Modeling Authors: Xiaoxiao Liang, Juyuan Zhang, Liming Pan, Linyuan Lü Published: 2026-04-14 Categories: cs.LG arXiv: https://arxiv.org/abs/2604.12806v1 Inferring latent interaction structures from observed dynamics is a fundamental inverse problem in many-body interacting systems. Most neural approaches rely on black-box surrogates over trainable graphs, achieving accuracy at the expense of mechanistic interpretability. Symbolic regression offers explicit dynamical equations and stronger inductive biases, but typically assumes known topology and a fixed function library. We propose \textbf{COSINE} (\textbf{C}o-\textbf{O}ptimization of \textbf{S}ymbolic \textbf{I}nteractions and \textbf{N}etwork \textbf{E}dges), a differentiable framework that jointly discovers interaction graphs and sparse symbolic dynamics. To overcome the limitations of fixed symbolic libraries, COSINE further incorporates an outer-loop large language model that adaptively prunes and expands the hypothesis space using feedback from the inner optimization loop. Experiments on synthetic systems and large-scale real-world epidemic data demonstrate robust structural recovery and compact, mechanism-aligned dynamical expressions. Code: https://anonymous.4open.science/r/COSINE-6D43. --- ### Image-to-Image Translation Framework Embedded with Rotation Symmetry Priors Authors: Feiyu Tan, Heran Yang, Qihong Duan, Kai Ye, Qi Xie, Deyu Meng Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12805v1 Image-to-image translation (I2I) is a fundamental task in computer vision, focused on mapping an input image from a source domain to a corresponding image in a target domain while preserving domain-invariant features and adapting domain-specific attributes. Despite the remarkable success of deep learning-based I2I approaches, the lack of paired data and unsupervised learning framework still hinder their effectiveness. In this work, we address the challenge by incorporating transformation symmetry priors into image-to-image translation networks. Specifically, we introduce rotation group equivariant convolutions to achieve rotation equivariant I2I framework, a novel contribution, to the best of our knowledge, along this research direction. This design ensures the preservation of rotation symmetry, one of the most intrinsic and domain-invariant properties of natural and scientific images, throughout the network. Furthermore, we conduct a systematic study on image symmetry priors on real dataset and propose a novel transformation learnable equivariant convolutions (TL-Conv) that adaptively learns transformation groups, enhancing symmetry preservation across diverse datasets. We also provide a theoretical analysis of the equivariance error of TL-Conv, proving that it maintains exact equivariance in continuous domains and provide a bound for the error in discrete cases. Through extensive experiments across a range of I2I tasks, we validate the effectiveness and superior performance of our approach, highlighting the potential of equivariant networks in enhancing generation quality and its broad applicability. Our code is available at https://github.com/tanfy929/Equivariant-I2I --- ### Generative Anonymization in Event Streams Authors: Adam T. Müller, Mihai Kocsis, Nicolaj C. Stache Published: 2026-04-14 Categories: cs.CV, cs.LG arXiv: https://arxiv.org/abs/2604.12803v1 Neuromorphic vision sensors offer low latency and high dynamic range, but their deployment in public spaces raises severe data protection concerns. Recent Event-to-Video (E2V) models can reconstruct high-fidelity intensity images from sparse event streams, inadvertently exposing human identities. Current obfuscation methods, such as masking or scrambling, corrupt the spatio-temporal structure, severely degrading data utility for downstream perception tasks. In this paper, to the best of our knowledge, we present the first generative anonymization framework for event streams to resolve this utility-privacy trade-off. By bridging the modality gap between asynchronous events and standard spatial generative models, our pipeline projects events into an intermediate intensity representation, leverages pretrained models to synthesize realistic, non-existent identities, and re-encodes the features back into the neuromorphic domain. Experiments demonstrate that our method reliably prevents identity recovery from E2V reconstructions while preserving the structural data integrity required for downstream vision tasks. Finally, to facilitate rigorous evaluation, we introduce a novel, synchronized real-world event and RGB dataset captured via precise robotic trajectories, providing a robust benchmark for future research in privacy-preserving neuromorphic vision. --- ### Efficiency of Proportional Mechanisms in Online Auto-Bidding Advertising Authors: Nguyen Kim Thang Published: 2026-04-14 Categories: cs.GT, cs.AI, cs.DS arXiv: https://arxiv.org/abs/2604.12799v2 The rise of automated bidding strategies in online advertising presents new challenges in designing and analyzing efficient auction mechanisms. In this paper, we focus on proportional mechanisms within the context of auto-bidding and study the efficiency of pure Nash equilibria, specifically the price of anarchy (PoA), under the liquid welfare objective. We first establish a tight PoA bound of 2 for the standard proportional mechanism. Next, we introduce a modified version with an alternative payment scheme that achieves a PoA bound of $1 + \frac{O(1)}{n-1}$ where $n \geq 2$ denotes the number of bidding agents. This improvement surpasses the existing PoA barrier of 2 and approaches full efficiency as the number of agents increases. Our methodology leverages duality and the Karush-Kuhn-Tucker (KKT) conditions from linear and convex programming. Due to its conceptual simplicity, our approach may offer broader applications for establishing PoA bounds. --- ### VFA: Relieving Vector Operations in Flash Attention with Global Maximum Pre-computation Authors: Yupeng Sun, Yanzhao Li, Zhiqiang Zou, Bai Du, Zhiyuan Zhang, Hui Dong, Gaoyige Fan, Hui Wang Published: 2026-04-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2604.12798v1 FlashAttention-style online softmax enables exact attention computation with linear memory by streaming score tiles through on-chip memory and maintaining a running maximum and normalizer. However, as attention kernels approach peak tensor-core/cube-core throughput on modern accelerators, non-matmul components of online softmax -- especially per-tile rowmax and rowsum reductions and rescale chains -- can become vector or SIMD limited and dominate latency. This paper revisits FlashAttention and proposes Vector Relieved Flash Attention (VFA), a hardware-friendly method that reduces rowmax-driven updates of the running maximum while retaining the online-softmax structure. VFA initializes the running maximum via a cheap approximation from key-block representations, reorders key-block traversal to prioritize high-impact sink and local blocks, and freezes the maximum for remaining blocks to avoid repeated reductions and rescaling. We further integrate VFA with block-sparse skipping methods such as BLASST to form Vector Relieved Sparse Attention (VSA), which reduces both block count and per-block overhead. Notably, VFA and VSA completely avoid the conditional rescale operation in the update stage used in FA4.0. Extensive evaluations on benchmarks including MMLU and MATH500, together with attention statistics, verify our design: (i) sink and local reordering stabilizes the running maximum early; (ii) simple Q and K block summaries fail due to intra-block heterogeneity; (iii) m-initialization is required when maxima appear in middle blocks. Overall, VFA and VSA efficiently alleviate online-softmax reduction bottlenecks without performance loss. Compared to the C16V32 baseline, C8V32, C4V32 and C4V16 achieve nearly two times speedup on modern hardware while hitting the vector bottleneck. With upcoming architecture improvements, C4V16 will deliver six times speedup by enhancing exponent capacity. --- ### OSC: Hardware Efficient W4A4 Quantization via Outlier Separation in Channel Dimension Authors: Zhiyuan Zhang, Yanzhao Li, Zhiqiang Zou, Bai Du, Yupeng Sun, Hui Dong, Hui Wang Published: 2026-04-14 Categories: cs.LG, cs.AI arXiv: https://arxiv.org/abs/2604.12782v1 While 4-bit quantization is essential for high-throughput deployment of Large Language Models, activation outliers often lead to significant accuracy degradation due to the restricted dynamic range of low-bit formats. In this paper, we systematically investigate the spatial distribution of outliers and demonstrate a token-persistent structural clustering effect, where high-magnitude outliers consistently occupy fixed channels across tokens. Building on this insight, we propose OSC, a hardware-efficient framework for outlier suppression. During inference, OSC executes a dual-path computation consisting of a low-precision 4-bit General Matrix Multiplication (GEMM) path and a high-precision 16-bit branch GEMM path. Specifically, OSC uses an offline group-wise strategy to identify the channels where outliers are located and then performs structured sub-tensor extraction to coalesce these scattered activation channels into a compact dense tensor online. This mechanism implements outlier protection through regularized and high-throughput GEMM operations, achieving a seamless fit with modern 4-bit micro-scaling hardware. Furthermore, for the inputs of W2 where outlier clustering is less pronounced, we integrate a fallback strategy to FP8. Evaluation on Qwen3-8B and Qwen3-30B restricts the average accuracy drop to 2.19 and 1.12 points, respectively. Notably, OSC is highly hardware-friendly, achieving a peak speedup of 1.78x over the W8A8 GEMM baseline on a modern AI accelerator. --- ### Fragile Reconstruction: Adversarial Vulnerability of Reconstruction-Based Detectors for Diffusion-Generated Images Authors: Haoyang Jiang, Mingyang Yi, Shaolei Zhang, Junxian Cai, Qingbin Liu, Xi Chen, Ju Fan Published: 2026-04-14 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.12781v1 Recently, detecting AI-generated images produced by diffusion-based models has attracted increasing attention due to their potential threat to safety. Among existing approaches, reconstruction-based methods have emerged as a prominent paradigm for this task. However, we find that such methods exhibit severe security vulnerabilities to adversarial perturbations; that is, by adding imperceptible adversarial perturbations to input images, the detection accuracy of classifiers collapses to near zero. To verify this threat, we present a systematic evaluation of the adversarial robustness of three representative detectors across four diverse generative backbone models. First, we construct adversarial attacks in white-box scenarios, which degrade the performance of all well-trained detectors. Moreover, we find that these attacks demonstrate transferability; specifically, attacks crafted against one detector can be transferred to others, indicating that adversarial attacks on detectors can also be constructed in a black-box setting. Finally, we assess common countermeasures and find that standard defense methods against adversarial attacks provide limited mitigation. We attribute these failures to the low signal-to-noise ratio (SNR) of attacked samples as perceived by the detectors. Overall, our results reveal fundamental security limitations of reconstruction-based detectors and highlight the need to rethink existing detection strategies. --- ### Efficient Adversarial Training via Criticality-Aware Fine-Tuning Authors: Wenyun Li, Zheng Zhang, Dongmei Jiang, Yaowei Wang, Xiangyuan Lan Published: 2026-04-14 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.12780v1 Vision Transformer (ViT) models have achieved remarkable performance across various vision tasks, with scalability being a key advantage when applied to large datasets. This scalability enables ViT models to exhibit strong generalization capabilities. However, as the number of parameters increases, the robustness of ViT models to adversarial examples does not scale proportionally. Adversarial training (AT), one of the most effective methods for enhancing robustness, typically requires fine-tuning the entire model, leading to prohibitively high computational costs, especially for large ViT architectures. In this paper, we aim to robustly fine-tune only a small subset of parameters to achieve robustness comparable to standard AT. To accomplish this, we introduce Criticality-Aware Adversarial Training (CAAT), a novel method that adaptively allocates resources to the most robustness-critical parameters, fine-tuning only selected modules. Specifically, CAAT efficiently identifies parameters that contribute most to adversarial robustness. It then leverages parameter-efficient fine-tuning (PEFT) to robustly adjust weight matrices where the number of critical parameters exceeds a predefined threshold. CAAT exhibits favorable generalization when scaled to larger vision transformer architectures, potentially paving the way for adversarial training at scale, e.g, compared with plain adversarial training, CAAT incurs only a 4.3% decrease in adversarial robustness while tuning approximately 6% of its parameters. Extensive experiments on three widely used adversarial learning datasets demonstrate that CAAT outperforms state-of-the-art lightweight AT methods with fewer trainable parameters. ---