# Research Papers — 2026-04 · bundle page 88 of 233 > Full-text markdown of papers #4351–4400 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/88 [← Previous bundle](https://aipod.dev/papers/archive/2026-04/page/87.md) [Next bundle →](https://aipod.dev/papers/archive/2026-04/page/89.md) ### SkillFlow:Benchmarking Lifelong Skill Discovery and Evolution for Autonomous Agents Authors: Ziao Zhang, Kou Shi, Shiting Huang, Avery Nie, Yu Zeng, Yiming Zhao, Zhen Fang, Qishen Su, Haibo Qiu, Wei Yang, Qingnan Ren, Shun Zou et al. Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17308v1 As the capability frontier of autonomous agents continues to expand, they are increasingly able to complete specialized tasks through plug-and-play external skills. Yet current benchmarks mostly test whether models can use provided skills, leaving open whether they can discover skills from experience, repair them after failure, and maintain a coherent library over time. We introduce SkillFlow, a benchmark of 166 tasks across 20 families in which task construction within each family follows a Domain-Agnostic Execution Flow (DAEF) that defines an agent workflow framework, allowing these tasks to share a consistent workflow. Agents are evaluated under an Agentic Lifelong Learning protocol in which they begin without skills, solve tasks sequentially within each family, externalize lessons through trajectory- and rubric-driven skill patches, and carry the updated library forward. Experiments reveal a substantial capability gap. For Claude Opus 4.6, lifelong skill evolution improves task success from 62.65% to 71.08% (+8.43 points). However, high skill usage does not necessarily imply high utility: Kimi K2.5 gains only +0.60 points despite 66.87% skill usage, while Qwen-Coder-Next reaches only a 44.58% task completion rate and still regresses relative to the vanilla setting. SkillFlow contributes a structured testbed for this direction and an in-depth empirical analysis of skill discovery, patching, transfer, and their failure modes under lifelong evaluation. --- ### Generalizable Face Forgery Detection via Separable Prompt Learning Authors: Enrui Yang, Baoyuan Wu, Yuezun Li Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17307v2 Detecting face forgeries using CLIP has recently emerged as a promising direction. However, most existing methods focus on adapting its visual encoder, leaving the potential of the textual encoder largely underexplored. In this paper, we propose Separable Prompt Learning (SePL) to better exploit the text modality, which further enhances the detection capacity. Specifically, SePL distills the forgery knowledge from CLIP via two separate learnable prompts, supported by a cross-modality alignment strategy and dedicated objectives. Extensive experiments demonstrate that our method achieves superior performance under both cross-dataset and cross-method evaluation. The code has been released at https://github.com/OUC-YER/SePL-DeepfakeDetection. --- ### Generalizable Face Forgery Detection via Separable Prompt Learning Authors: Enrui Yang, Yuezun Li Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17307v1 Detecting face forgeries using CLIP has recently emerged as a promising and increasingly popular research direction. Owing to its rich visual knowledge acquired through large-scale pretraining, most existing methods typically rely on the visual encoder of CLIP, while paying limited attention to the text modality. Given the instructive nature of the text modality, we posit that it can be leveraged to instruct Deepfake detection with meticulous design. Accordingly, we shift the focus from the visual modality to the text modality and propose a new Separable Prompt Learning strategy (SePL) that enables CLIP to serve as an effective face forgery detector. The core idea of SePL is to disentangle forgery-specific and forgery-irrelevant information in images via two types of prompt learning, with the former enhancing detection. To achieve this disentangle, we describe a cross-modality alignment strategy and a set of dedicated objectives. Extensive experiments demonstrate that, with this simple adaptation, our method achieves competitive and even superior performance compared to other methods under both cross-dataset and cross-method evaluation, highlighting its strong generalizability. The codes have been released at https://github.com/OUC-YER/SePL-DeepfakeDetection --- ### The First Challenge on Mobile Real-World Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview Authors: Jiatong Li, Zheng Chen, Kai Liu, Jingkai Wang, Zihan Zhou, Xiaoyang Liu, Libo Zhu, Jue Gong, Radu Timofte, Yulun Zhang, Congyu Wang, Zihao Wang et al. Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17306v1 This paper provides a review of the NTIRE 2026 challenge on mobile real-world image super-resolution, highlighting the proposed solutions and the resulting outcomes. The challenge aims to recover high-resolution (HR) images from low-resolution (LR) counterparts generated through unknown degradations with a x4 scaling factor while ensuring the models remain executable on mobile devices. The objective is to develop effective and efficient network designs or solutions that achieve state-of-the-art real-world image super-resolution performance. The track of the challenge evaluates performance using a weighted combination of image quality assessment (IQA) score and speedup ratios. The competition attracted 108 registrants, with 16 teams achieving a valid score in the final ranking. This collaborative effort advances the performance of mobile real-world image super-resolution while offering an in-depth overview of the latest trends in the field. --- ### Efficient Test-Time Scaling via Temporal Reasoning Aggregation Authors: Jiakun Li, Xingwei He, Kefan Li, Hongzheng Chai, Hongyue Yu, Yuan Yuan Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17304v1 Test-time scaling improves the reasoning performance of large language models but often results in token-inefficient overthinking, where models continue reasoning beyond what is necessary for a correct answer. Existing dynamic early-exit methods typically rely on single-step confidence signals, which are often unreliable for detecting reasoning convergence in multi-step settings. To mitigate this limitation, we propose TRACE, a training-free framework for efficient test-time scaling that determines when to terminate reasoning based on temporal aggregation of multi-step evidence rather than instantaneous signals. TRACE detects reasoning convergence over time by aggregating two complementary signals across recent reasoning steps: answer consistency, capturing the persistence of predicted answers, and confidence trajectory, modeling the temporal evolution of model confidence. Benefiting from these two factors, TRACE can accurately determine whether the reasoning process has converged, thereby promptly halting inference and effectively avoiding redundant reasoning steps. Extensive experiments on multiple challenging benchmarks show that TRACE reduces reasoning token usage by 25-30% on average while maintaining accuracy within 1-2% of full-length reasoning, consistently outperforming existing dynamic reasoning methods. --- ### RoTRAG: Rule of Thumb Reasoning for Conversation Harm Detection with Retrieval-Augmented Generation Authors: Juhyeon Lee, Wonduk Seo, Junseo Koh, Seunghyun Lee, Haihua Chen, Yi Bu Published: 2026-04-19 Categories: cs.CL, cs.AI, cs.HC, cs.IR, cs.LG arXiv: https://arxiv.org/abs/2604.17301v2 Detecting harmful content in multi turn dialogue requires reasoning over the full conversational context rather than isolated utterances. However, most existing methods rely mainly on models internal parametric knowledge, without explicit grounding in external normative principles. This often leads to inconsistent judgments in socially nuanced contexts, limited interpretability, and redundant reasoning across turns. To address this, we propose RoTRAG, a retrieval augmented framework that incorporates concise human written moral norms, called Rules of Thumb (RoTs), into LLM based harm assessment. For each turn, RoTRAG retrieves relevant RoTs from an external corpus and uses them as explicit normative evidence for turn level reasoning and final severity classification. To improve efficiency, we further introduce a lightweight binary routing classifier that decides whether a new turn requires retrieval grounded reasoning or can reuse existing context. Experiments on ProsocialDialog and Safety Reasoning Multi Turn Dialogue show that RoTRAG consistently improves both harm classification and severity estimation over competitive baselines, with an average relative gain of around 40% in F1 across benchmark datasets and an average relative reduction of 8.4% in distributional error, while reducing redundant computation without sacrificing performance. --- ### Chaos-Enhanced Prototypical Networks for Few-Shot Medical Image Classification Authors: Chinthakuntla Meghan Sai, Murarisetty V Sai Kartheek, Sita Devi Bharatula, Karthik Seemakurthy Published: 2026-04-19 Categories: eess.IV, cs.AI, cs.CV arXiv: https://arxiv.org/abs/2604.17300v1 The scarcity of labeled clinical data in oncology makes Few-Shot Learning (FSL) a critical framework for Computer Aided Diagnostics, but we observed that standard Prototypical Networks often struggle with the "prototype instability" caused by morphological noise and high intra-class variance in brain tumor scans. Our work attempts to minimize this by integrating a non-linear Logistic Chaos Module into a fine-tuned ResNet-18 backbone creating the Chaos-Enhanced ProtoNet(CE-ProtoNet). Using the deterministic ergodicity of the logistic chaos map we inject controlled perturbations into support features during episodic training-essentially for "stress testing" the embedding space. This process makes the model to converge on noise-invariant representations without increasing computational overhead. Testing this on a 4-way 5-shot brain tumor classification task, we found that a 15% chaotic injection level worked efficiently to stabilize high-dimensional clusters and reduce class dispersion. Our method achieved a peak test accuracy of 84.52%, outperforming standard ProtoNet. Our results suggest the idea of using chaotic perturbation as an efficient, low-overhead regularization tool, for the data-scarce regimes. --- ### Cat-DPO: Category-Adaptive Safety Alignment Authors: Tiankai Yang, Yi Nian, Xinyuan Li, Ruiyao Xu, Henry Peng Zou, Kaize Ding, Xiyang Hu, Yan Liu, Yue Zhao Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17299v3 Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a single scalar that is applied uniformly to every preference pair. The result is a model that looks safe on average but stays relatively unsafe on a minority of harm categories. We cast safety alignment as a per-category constrained optimization problem and derive Cat-DPO, a direct-preference-optimization algorithm with a separate adaptive safety margin for each harm category. The margin tightens when the model still produces unsafe responses on a category and relaxes once the model catches up, so the training signal tracks each category's current difficulty rather than averaging under one global rate. Across two LLM backbones and six preference-learning baselines, Cat-DPO improves aggregate helpfulness and harmlessness and compresses per-category safety variance and the best-to-worst gap, offering a drop-in per-category refinement of direct preference safety alignment. --- ### Cat-DPO: Category-Adaptive Safety Alignment Authors: Tiankai Yang, Yi Nian, Xinyuan Li, Ruiyao Xu, Kaize Ding, Yue Zhao Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17299v2 Aligning large language models with human preferences must balance two competing goals: responding helpfully to legitimate requests and reliably refusing harmful ones. Most preference-based safety alignment methods collapse safety into a single scalar that is applied uniformly to every preference pair. The result is a model that looks safe on average but stays relatively unsafe on a minority of harm categories. We cast safety alignment as a per-category constrained optimization problem and derive Cat-DPO, a direct-preference-optimization algorithm with a separate adaptive safety margin for each harm category. The margin tightens when the model still produces unsafe responses on a category and relaxes once the model catches up, so the training signal tracks each category's current difficulty rather than averaging under one global rate. Across two LLM backbones and six preference-learning baselines, Cat-DPO improves aggregate helpfulness and harmlessness and compresses per-category safety variance and the best-to-worst gap, offering a drop-in per-category refinement of direct preference safety alignment. --- ### Frequency-guided Multi-level Reasoning for Scene Graph Generation in Video Authors: Chenxing Li, Yiping Duan, Xiaoming Tao Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17298v1 Video Scene Graph Generation aims to obtain structured semantic representations of objects and their relationships in videos for high-level understanding. However, existing methods still have limitations in handling long-tail distributions. This paper proposes the Frequency-guided Relational Multi-level Reasoning (FReMuRe) model, which enhances the modeling ability of long-tail relationships from a mechanism perspective. We introduce relation-specific branches to deal gradient conflicts, yielding more balanced and tail-aware learning. And we design a frequency-aware dual-branch predicate embedding network to model high-frequency and low-frequency relationships separately and improve the recall rate of tail classes through gated fusion. Meanwhile, we propose two types of interchangeable relation classification heads: Bayesian Head for uncertainty estimation and new Gaussian Mixture Model Head to enhance intra-class diversity. Experimental results show that FReMuRe significantly improves the recall rate of long-tail relationships and overall reasoning robustness on the Action Genome dataset. --- ### CRISP: Compressing Redundancy in Chain-of-Thought via Intrinsic Saliency Pruning Authors: Yangsong Lan, Hongliang Dai, Piji Li Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17297v1 Long Chain-of-Thought (CoT) reasoning is pivotal for the success of recent reasoning models but suffers from high computational overhead and latency. While prior works attempt to compress CoT via external compressor, they often fail to align with the model's internal reasoning dynamics, resulting in the loss of critical logical steps. This paper presents \textbf{C}ompressing \textbf{R}edundancy in Chain-of-Thought via \textbf{I}ntrinsic \textbf{S}aliency \textbf{P}runing (\textbf{CRISP}), a framework that compresses CoT by exploiting the model's intrinsic saliency. Our analysis reveals a distinct phenomenon: the reasoning termination token \texttt{(object Object)} acts as an information anchor, where its attention pattern effectively demarcates essential reasoning from redundancy. Based on this finding, we design a policy that utilizes these intrinsic attention signals to guide atomic compression operations. In contrast to coarse-grained pruning strategies, CRISP strategically distills the reasoning chain to maximize information density while preserving logical coherence. Empirical results across various backbone models and mathematical datasets demonstrate that CRISP achieves a 50-60% reduction in token count without compromising accuracy, effectively mitigating the efficiency bottleneck of long-context reasoning. We open-source our implementation to facilitate further research in efficient reasoning. --- ### LLaTiSA: Towards Difficulty-Stratified Time Series Reasoning from Visual Perception to Semantics Authors: Yueyang Ding, HaoPeng Zhang, Rui Dai, Yi Wang, Tianyu Zong, Kaikui Liu, Xiangxiang Chu Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17295v1 Comprehensive understanding of time series remains a significant challenge for Large Language Models (LLMs). Current research is hindered by fragmented task definitions and benchmarks with inherent ambiguities, precluding rigorous evaluation and the development of unified Time Series Reasoning Models(TSRMs). To bridge this gap, we formalize Time Series Reasoning (TSR) via a four-level taxonomy of increasing cognitive complexity. We introduce HiTSR, a hierarchical time series reasoning dataset comprising 83k samples with diverse task combinations and verified Chain-of-Thought (CoT) trajectories. Leveraging HiTSR, we propose LLaTiSA, a strong TSRM that integrates visualized patterns with precision-calibrated numerical tables to enhance the temporal perception of Vision-Language Models (VLMs). Through a multi-stage curriculum fine-tuning strategy, LLaTiSA achieves superior performance and exhibits robust out-of-distribution generalization across diverse TSR tasks and real-world scenarios. Our code is available at https://github.com/RainingNovember/LLaTiSA. --- ### Beyond "I Don't Know": Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty Authors: Jingyi Ren, Ante Wang, Yunghwei Lai, Xiaolong Wang, Linlu Gong, Weitao Li, Weizhi Ma, Yang Liu Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17293v1 Reliable Large Language Models (LLMs) should abstain when confidence is insufficient. However, prior studies often treat refusal as a generic "I don't know'', failing to distinguish input-level ambiguity (data uncertainty) from capability limitations (model uncertainty). This lack of distinction limits downstream action decisions like requesting clarification or invoking external tools. In this work, we introduce UA-Bench, a benchmark of over 3,500 questions drawn from six datasets spanning knowledge-intensive and reasoning-intensive tasks, designed to evaluate explicit uncertainty attribution. An evaluation of 18 frontier LLMs shows that even state-of-the-art models struggle to reliably discriminate between data uncertainty and model uncertainty, and that high answer accuracy does not necessarily imply strong uncertainty attribution ability. To narrow this gap, we propose a lightweight data synthesis and reinforcement learning strategy. Experiments on both Qwen3-4B-Instruct-2507 and Qwen3-8B in thinking mode show that the proposed method improves uncertainty attribution while preserving answer accuracy. Our code and data are publicly available now. --- ### Probabilistic Programs of Thought Authors: Poorva Garg, Renato Lui Geh, Daniel Israel, Todd Millstein, Kyle Richardson, Guy Van den Broeck Published: 2026-04-19 Categories: cs.CL, cs.AI, cs.PL arXiv: https://arxiv.org/abs/2604.17290v1 LLMs are widely used for code generation and mathematical reasoning tasks where they are required to generate structured output. They either need to reason about code, generate code for a given specification, or reason using programs of thought. The typical approach to code generation is to prompt the model and generate samples until an appropriate program is obtained. Within this process, sampling $n$ programs from the language model requires $n$ GPU compute-intensive generations which becomes prohibitively expensive for larger values of $n$. In this work, we address this limitation by exposing the LLM's distribution within the generated programs themselves. We propose a novel test-time framework we dub probabilistic programs of thought to obtain more samples from the model with fewer LLM generations. Given a program generated by a model and the associated next-token probabilities, we build a probabilistic program that compactly represents exponentially many deterministic programs. Since performing probabilistic reasoning in this probabilistic program is much cheaper, our approach allows sampling new programs without any additional GPU compute and little CPU overhead. We instantiate our approach on benchmarks for code generation, code understanding and mathematical reasoning and report improvements in performance with fewer generations from the LLM. --- ### REALM: Reliable Expertise-Aware Language Model Fine-Tuning from Noisy Annotations Authors: Sajjad Ghiasvand, Mark Beliaev, Mahnoosh Alizadeh, Ramtin Pedarsani Published: 2026-04-19 Categories: cs.LG arXiv: https://arxiv.org/abs/2604.17289v3 Supervised fine-tuning of large language models relies on human-annotated data, yet annotation pipelines routinely involve multiple crowdworkers of heterogeneous expertise. Standard practice aggregates labels via majority vote or simple averaging, discarding annotator identity and causing the model to absorb the errors of unreliable annotators into its parameters. We propose REALM, which jointly learns the model parameters and a scalar expertise value for each annotator, entirely unsupervised and requiring nothing beyond annotator identity. The key idea is to model each observed label as a mixture between the model's prediction and a uniform random guess, weighted by the annotator's learned expertise. REALM applies to any task with a fixed label set, and extends to multiple tasks via a learned expertise matrix. On four text-classification datasets with \emph{real} crowdsourced annotations, REALM is the best method in all $12$ configurations of the three heterogeneous-annotator datasets, improving on the strongest applicable baseline, including majority vote and Dawid--Skene aggregation, by $+2.9$ points on average. On five question answering benchmarks with simulated noisy labels, it outperforms naive noisy fine-tuning in $152$ of $162$ configurations, by $+5.0$ points on average, with gains that grow with model capacity. The learned expertise additionally recovers annotator reliability without ever observing it. Our code is available at https://github.com/sajjad-ucsb/REALM --- ### REALM: Reliable Expertise-Aware Language Model Fine-Tuning from Noisy Annotations Authors: Sajjad Ghiasvand, Mark Beliaev, Mahnoosh Alizadeh, Ramtin Pedarsani Published: 2026-04-19 Categories: cs.LG arXiv: https://arxiv.org/abs/2604.17289v2 Supervised fine-tuning of large language models relies on human-annotated data, yet annotation pipelines routinely involve multiple crowdworkers of heterogeneous expertise. Standard practice aggregates labels via majority vote or simple averaging, discarding annotator identity and causing the model to absorb the errors of unreliable annotators directly into its parameters. We propose REALM, a method that jointly learns the model parameters and a scalar expertise value for each annotator entirely unsupervised, requiring no supervision beyond annotator identity. The key idea is to model each observed label as a mixture between the model's prediction and a uniform random guess, weighted by the annotator's learned expertise. We extend REALM to a multi-task setting via a learned expertise matrix that captures per-annotator reliability across tasks. We evaluate on five question answering benchmarks, fine-tuning three sizes of Flan-T5 under simulated noisy annotations. The proposed algorithm consistently outperforms the naive noisy SFT in the large majority of single- and multi-task settings, across datasets, model sizes, and noise types, with accuracy improvements of up to $50\%$ in the most adversarial regime and gains that grow with model capacity. --- ### Clover: A Neural-Symbolic Agentic Harness with Stochastic Tree-of-Thoughts for Verified RTL Repair Authors: Zizhang Luo, Yansong Xu, Runlin Guo, Fan Cui, Kexing Zhou, Mile Xia, Hongyuan Hou, Yuhao Luo, Yun Liang Published: 2026-04-19 Categories: cs.AR, cs.AI arXiv: https://arxiv.org/abs/2604.17288v1 RTL program repair remains a critical bottleneck in hardware design and verification. Traditional automatic program repair (APR) methods rely on predefined templates and synthesis, limiting their bug coverage. Large language models (LLMs) and coding agents based on them offer flexibility but suffer from randomness and context corruption when handling long RTL code and waveforms. We present Clover, a neural-symbolic agentic harness that orchestrates RTL repair as a structured search over code manipulations to explore a validated solution for the bug. Recognizing that different repair operations favor distinct strategies, Clover dynamically dispatches tasks to specialized LLM agents or symbolic solvers. At its core, Clover introduces stochastic tree-of-thoughts, a test-time scaling mechanism that manages the main agent's context as a search tree, balancing exploration and exploitation for reliable outcomes. An RTL-specific toolbox further empowers agents to interact with the debugging environment. Evaluated on the RTL-repair benchmark, Clover fixes 96.8% of bugs within a fixed time limit, covering 94% and 63% more bugs than both pure traditional and LLM-based baselines, respectively, while achieving an average pass@1 rate of 87.5%, demonstrating high reliability and effectiveness. --- ### Spectral Forensics of Diffusion Attention Graphs for Copy-Move Forgery Detection Authors: H. M. Shadman Tabib, Tasriad Ahmed Tias, Nafis Tahmid Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17287v1 Copy-move forgery, where a region within an image is duplicated to hide or fabricate content, remains a persistent threat to visual media integrity. We introduce GraphSpecForge, a training-free framework that detects copy-move forgery by analysing the spectral structure of attention graphs from a pretrained Stable Diffusion U-Net. Our central insight is that copy-move manipulation induces approximate subgraph duplication in the self-attention graph, leading to measurable spectral redistribution in the normalized graph Laplacian. We formalise this link with perturbation-based arguments and build an image-level anomaly detector using Wasserstein distances between per-image Laplacian spectra and an authentic reference distribution. We evaluate GraphSpecForge on four copy-move benchmarks without forgery-specific retraining. On RecodAI-LUC (5,128 images), our best configuration achieves AUROC = 0.606 (95% CI: 0.580-0.638; permutation p = 0.005), and the normalized Laplacian outperforms raw attention spectra by +0.057 AUROC. On MICC-F220, CoMoFoD, and COVERAGE, the same pipeline attains AUROCs of 0.752, 0.774, and 0.673, respectively; on CoMoFoD it also reaches AUPRC = 0.833, balanced accuracy = 0.712, MCC = 0.499, and TPR@1%FPR = 32.5%. Additional ablation and falsification experiments confirm the signal's specificity and sensitivity to manipulation strength, while null-graph controls rule out trivial-statistic explanations. --- ### Depth Adaptive Efficient Visual Autoregressive Modeling Authors: Chunliang Li, Tianze Cao, Sanyuan Zhao Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17286v1 Visual Autoregressive (VAR) modeling inefficiently applies a fixed computational depth to each position when generating high-resolution images. While existing methods accelerate inference by pruning tokens using frequency maps, their binary hard-pruning approach is fundamentally limited and fails to improve quality even with better frequency estimation. Observing that VAR models possess significant depth redundancy, we propose a paradigm shift from pruning entire tokens to adaptively allocating per-token computational depth. To this end, we introduce DepthVAR, a training-free framework that dynamically allocates computation. It integrates an adaptive depth scheduler, which assigns computational depth via a cyclic rotated schedule for balanced, non-static refinement, with a dynamic inference process that translates these depths into layer-major masks, selectively applies transformer blocks, and blends the resulting codes to ensure each token's influence is proportional to its processing depth. Extensive experiments show that DepthVAR achieves 2.3$\times$-3.1$\times$ acceleration with minimal quality loss, offering a competitive compute-performance trade-off compared to existing hard-pruning approaches. Code is available at https://github.com/STOVAGtz/DepthVAR --- ### HalluClear: Diagnosing, Evaluating and Mitigating Hallucinations in GUI Agents Authors: Chao Jin, Wenkui Yang, Hao Sun, Yuqi Liao, Qianyi Jiang, Kai Zhou, Jie Cao, Ran He, Huaibo Huang Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17284v1 While progress in GUI agents has been largely driven by industrial-scale training, ungrounded hallucinations often trigger cascading failures in real-world deployments.Unlike general VLM domains, the GUI agent field lacks a hallucination-focused suite for fine-grained diagnosis, reliable evaluation, and targeted mitigation.To bridge this gap, we introduce HalluClear, a comprehensive suite for hallucination mitigation in GUI agents as a complement to computation-intensive scaling. HalluClear comprises: (1) a GUI-specific hallucination taxonomy derived from empirical failure analysis; (2) a calibrated three-stage evaluation workflow which enhances VLM-as-a-judge reliability via expert-annotated benchmarking and ensemble credibility estimation; and (3) a mitigation scheme based on closed-loop structured reasoning, enabling lightweight continual post-training with cold-start initialization for both generalist and GUI-specialist agents. Experiments across representative agents and public benchmarks demonstrate that post-training on only 9K samples within our suite can significantly reduce hallucinations, thereby improving grounding and action fidelity, offering a compute-efficient pathway to robust GUI automation. --- ### HorizonBench: Long-Horizon Personalization with Evolving Preferences Authors: Shuyue Stella Li, Bhargavi Paranjape, Kerem Oktar, Zhongyao Ma, Gelin Zhou, Lin Guan, Na Zhang, Sem Park, Lin Chen, Diyi Yang, Yulia Tsvetkov, Asli Celikyilmaz Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17283v1 User preferences evolve across months of interaction, and tracking them requires inferring when a stated preference has been changed by a subsequent life event. We define this problem as long-horizon personalization and observe that progress on it is limited by data availability and measurement, with no existing resource providing both naturalistic long-horizon interactions and the ground-truth provenance needed to diagnose why models fail. We introduce a data generator that produces conversations from a structured mental state graph, yielding ground-truth provenance for every preference change across 6-month timelines, and from it construct HorizonBench, a benchmark of 4,245 items from 360 simulated users with 6-month conversation histories averaging ~4,300 turns and ~163K tokens. HorizonBench provides a testbed for long-context modeling, memory-augmented architectures, theory-of-mind reasoning, and user modeling. Across 25 frontier models, the best model reaches 52.8% and most score at or below the 20% chance baseline. When these models err on evolved preferences, over a third of the time they select the user's originally stated value without tracking the updated user state. This belief-update failure persists across context lengths and expression explicitness levels, identifying state-tracking capability as the primary bottleneck for long-horizon personalization. --- ### MedPRMBench: A Fine-grained Benchmark for Process Reward Models in Medical Reasoning Authors: Lingyan Wu, Xiang Zheng, Weiqi Zhai, Wei Wang, Xuan Ren, Zifan Zhang, Hu Wei, Bing Zhao Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17282v1 Process-Level Reward Models (PRMs) are essential for guiding complex reasoning in large language models, yet existing PRM benchmarks cover only general domains such as mathematics, failing to address medical reasoning -- which is uniquely characterized by safety criticality, knowledge intensity, and diverse error patterns. Without a reliable medical PRM evaluation framework, we cannot quantify models' error detection capabilities in clinical reasoning, leaving their safety in real-world healthcare applications unverified. We propose MedPRMBench, the first process-level reward model benchmark for the medical domain. Built through a three-phase pipeline based on Clinical Reasoning Blueprints (CRBs), MedPRMBench systematically generates high-quality evaluation data from seven medical QA sources, covering 14 fine-grained error types across three categories (Simplicity, Soundness, and Sensitivity) with the first 4-level severity grading system to quantify clinical impact. The benchmark comprises 6{,}500 questions with 13{,}000 reasoning chains and 113{,}910 step-level labels, plus 6{,}879 questions for training. Our medical PRM baseline achieves an 87.1\% overall PRMScore -- substantially surpassing all baselines -- and serves as a plug-and-play verifier that improves downstream medical QA accuracy by 3.2--6.7 percentage points. Systematic evaluation spanning proprietary frontier models, open-source reasoning models, and medical-specialized models reveals critical weaknesses in current models' medical reasoning error detection capabilities, providing clear directions for future PRM improvement. --- ### PestVL-Net: Enabling Multimodal Pest Learning via Fine-grained Vision-Language Interaction Authors: Xueheng Li, Tao Hu, Ke Cao, Runsheng Qi, Huixin Zhang, Rui Li, Jie Zhang, Chengjun Xie Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17278v1 Effective pest recognition and management are crucial for sustainable agricultural development. However, collecting pest data in real scenarios is often challenging. Compared to other domains, pests exhibit a wide variety of species with complex and diverse morphological characteristics. Existing techniques struggle to effectively model the key visual and high-level semantic features of pests in a fine-grained manner. These limitations hinder the practical application of such methods in real agricultural scenarios. To address these critical challenges, we present a synergistic approach that integrates PestVL-Net, a novel vision-language framework, with two multi-species pest datasets to facilitate fine-grained pest learning. The visual pathway of PestVL-Net utilizes the Recurrent Weighted Key Value (RWKV) architecture, incorporating a saliency-guided adaptive window partitioning scheme to effectively model the fine-grained visual characteristics of pests. Concurrently, the linguistic component generates precise pest semantic descriptions by leveraging Multimodal Large Language Models (MLLMs) priors, critically informed by agricultural expert knowledge and structured via multimodal Chain-of-Thought (CoT) reasoning. The deep fusion of these complementary visual and textual representations enables fine-grained multimodal pest learning. Extensive experimental evaluations on multiple pest datasets validate the superior performance of PestVL-Net, highlighting its potential for effective real-world pest management. --- ### Fully Analog Resonant Recurrent Neural Network via Metacircuit Authors: Zixin Zhou, Tianxi Jiang, Menglong Yang, Zhihua Feng, Qingbo He, Shiwu Zhang Published: 2026-04-19 Categories: cs.LG, cs.AI, cs.ET, physics.app-ph arXiv: https://arxiv.org/abs/2604.17277v1 Physical neural networks offer a transformative route to edge intelligence, providing superior inference speed and energy efficiency compared to conventional digital architectures. However, realizing scalable, end-to-end, fully analog recurrent neural networks for temporal information processing remains challenging due to the difficulty of faithfully mapping trained network models onto physical hardware. Here we present a fully analog resonant recurrent neural network (R$^2$NN) implemented via a metacircuit architecture composed of coupled electrical local resonators. A reformulated mechanical-electrical analogy establishes a direct mapping between the R$^2$NN model and metacircuit elements, enabling accurate physical implementation of trained neural network parameters. By integrating jointly trainable global resistive coupling and local resonances, which generate effective frequency-dependent negative resistances, the architecture shapes an impedance landscape that steers currents along frequency-selective pathways. This mechanism enables direct extraction of discriminative spectral features, facilitating real-time temporal classification of raw analog inputs while bypassing analog-to-digital conversion. We demonstrate the cross-domain versatility of this framework using integrated hardware for tactile perception, speech recognition, and condition monitoring. This work establishes a scalable, fully analog paradigm for intelligent temporal processing and paves the way for low-latency, resource-efficient physical neural hardware for edge intelligence. --- ### Instinct vs. Reflection: Unifying Token and Verbalized Confidence in Multimodal Large Models Authors: Yunkai Dang, Yifan Jiang, Yizhu Jiang, Anqi Chen, Wenbin Li, Yang Gao Published: 2026-04-19 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.17274v1 Multimodal Large Language Models (MLLMs) have demonstrated exceptional capabilities in various perception and reasoning tasks. Despite this success, ensuring their reliability in practical deployment necessitates robust confidence estimation. Prior works have predominantly focused on text-only LLMs, often relying on computationally expensive self-consistency sampling. In this paper, we extend this to multimodal settings and conduct a comprehensive evaluation of MLLMs' response confidence estimation. Our analysis reveals a significant instinct-reflection misalignment: the model's implicit token-level support frequently diverges from its verbal self-assessment confidence. To address this misalignment, we propose a monotone confidence fusion framework to merge dual-channel signals and cross-channel consistency to estimate correctness. Subsequently, an order-preserving mean alignment step is applied to correct global bias, which improves calibration while preserving the risk-coverage trade-off for selective prediction. Experiments on diverse open-source and closed-source MLLMs show that our method consistently yields more reliable confidence estimates and improves both calibration and failure prediction. Code will be available at https://github.com/Yunkaidang/Instinct-vs.-Reflection. --- ### The Continuity Layer: Why Intelligence Needs an Architecture for What It Carries Forward Authors: Samuel Sameer Tanguturi Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17273v1 The most important architectural problem in AI is not the size of the model but the absence of a layer that carries forward what the model has come to understand. Sessions end. Context windows fill. Memory APIs return flat facts that the model has to reinterpret from scratch on every read. The result is intelligence that is powerful per session and amnesiac across time. This position paper argues that the layer which fixes this, the continuity layer, is the most consequential piece of infrastructure the field has not yet built, and that the engineering work to build it has begun in public. The formal evaluation framework for the property described here is the ATANT benchmark (arXiv:2604.06710), published separately with evaluation results on a 250-story corpus; a companion paper (arXiv:2604.10981) positions this framework against existing memory, long-context, and agentic-memory benchmarks. The paper defines continuity as a system property with seven required characteristics, distinct from memory and from retrieval; describes a storage primitive (Decomposed Trace Convergence Memory) whose write-time decomposition and read-time reconstruction produce that property; maps the engineering architecture to the theological pattern of kenosis and the symbolic pattern of Alpha and Omega, and argues this mapping is structural rather than metaphorical; proposes a four-layer development arc from external SDK to hardware node to long-horizon human infrastructure; examines why the physics limits now constraining the model layer make the continuity layer newly consequential; and argues that the governance architecture (privacy implemented as physics rather than policy, founder-controlled class shares on non-negotiable architectural commitments) is inseparable from the product itself. --- ### HopRank: Self-Supervised LLM Preference-Tuning on Graphs for Few-Shot Node Classification Authors: Ziqing Wang, Kaize Ding Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17271v1 Node classification on text-attributed graphs (TAGs) is a fundamental task with broad applications in citation analysis, social networks, and recommendation systems. Current GNN-based approaches suffer from shallow text encoding and heavy dependence on labeled data, limiting their effectiveness in label-scarce settings. While large language models (LLMs) naturally address the text understanding gap with deep semantic reasoning, existing LLM-for-graph methods either still require abundant labels during training or fail to exploit the rich structural signals freely available in graph topology. Our key observation is that, in many real-world TAGs, edges predominantly connect similar nodes under the homophily principle, meaning graph topology inherently encodes class structure without any labels. Building on this insight, we reformulate node classification as a link prediction task and present HopRank, a fully self-supervised LLM-tuning framework for TAGs. HopRank constructs preference data via hierarchical hop-based sampling and employs adaptive preference learning to prioritize informative training signals without any class labels. At inference, nodes are classified by predicting their connection preferences to labeled anchors, with an adaptive early-exit voting scheme to improve efficiency. Experiments on three TAG benchmarks show that HopRank matches fully-supervised GNNs and substantially outperforms prior graph-LLM methods, despite using zero labeled training data. --- ### Understanding U.S. Users' Security and Privacy Transparency Needs for Consumer-Facing Generative AI Authors: Jiaxun Cao, Yu Dong, Chunxi Zhan, Rithvik Neti, Sai Teja Peddinti, Pardis Emami-Naeini Published: 2026-04-19 Categories: cs.HC, cs.AI, cs.CR, cs.CY arXiv: https://arxiv.org/abs/2604.17270v2 Users increasingly rely on consumer-facing generative AI (GenAI) for tasks ranging from everyday needs to sensitive use cases. Yet, it remains unclear whether and how existing security and privacy (S&P) communications in GenAI tools shape users' adoption decisions and experiences. Understanding how users seek, interpret, and evaluate S&P information is critical for designing usable transparency that users can trust and act on. We conducted semi-structured interviews and design sessions with 21 U.S. GenAI users. Our findings suggest that available S&P information rarely drove initial adoption in practice, as participants often perceived it as incomplete, ineffective, or not credible. Instead, they relied on rough proxies (e.g., popularity) to infer S&P practices. After adoption, S&P uncertainty constrained participants' willingness to use GenAI tools, especially for high-stakes purposes, and, in some cases, contributed to discontinued use. Participants therefore called for transparency that supports decisions and actions through trustworthy information (e.g., independent evaluations) and usable interfaces (e.g., on-demand disclosure). We categorize participants' desired design practices into five dimensions to facilitate systematic future investigation into best practices. We conclude with recommendations for researchers, designers, and policymakers to improve S&P transparency in consumer-facing GenAI. --- ### Fractal Characterization of Low-Correlation Signals in AI-Generated Image Detection Authors: Wenwei Xie, Jie Yin, Lu Ma, Xuansong Zhang, Wenjing Zhang Published: 2026-04-19 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.17268v1 AI-generated imagery has reached near-photorealistic fidelity, yet this technology poses significant threats to information security and societal trust. Existing deepfake detection methods often exhibit limited robustness in open-world scenarios. To address this limitation, this paper investigates intrinsic discrepancies between synthetic and authentic images from a signal-level perspective. Our analysis reveals that low-correlation signals serve as distinctive markers for differentiating AI-generated imagery from real photographs. Building on this insight, we introduce a novel method for quantifying these signals based on fractal theory. By analyzing the fractal characteristics of low-correlation signals, our method effectively captures the subtle statistical anomalies inherent to the synthesis process. Extensive experimental results demonstrate the method's robustness and superior detection performance. This work emphasizes the need to shift research focus to a new signal-level direction for deepfake detection. Theoretically, this proposed approach is not limited to face image identification but can be applied to all AI-generated image detection tasks. This study provides a new research direction for deepfake detection. --- ### Rectification Difficulty and Optimal Sample Allocation in LLM-Augmented Surveys Authors: Zikun Ye, Hema Yoganarasimhan Published: 2026-04-19 Categories: cs.AI, stat.AP arXiv: https://arxiv.org/abs/2604.17267v2 Large Language Models can generate synthetic survey responses at low cost, but their accuracy varies unpredictably across questions. We study the design problem of allocating a fixed budget of human respondents across estimation tasks when cheap LLM predictions are available for every task. Our framework combines three components. First, building on Prediction-Powered Inference, we characterize a question-specific rectification difficulty that governs how quickly the estimator's variance decreases with human sample size. Second, we derive a closed-form optimal allocation rule that directs more human labels to tasks where the LLM is least reliable. Third, since rectification difficulty depends on unobserved human responses for new surveys, we propose a meta-learning approach, trained on historical data, that predicts it for entirely new tasks without pilot data. The framework extends to general M-estimation, covering regression coefficients and multinomial logit partworths for conjoint analysis. We validate the framework on two datasets spanning different domains, question types, and LLMs, showing that our approach captures 61-79% of the theoretically attainable efficiency gains, achieving 11.4% and 10.5% MSE reductions without requiring any pilot human data for the target survey. --- ### Rethinking Meeting Effectiveness: A Benchmark and Framework for Temporal Fine-grained Automatic Meeting Effectiveness Evaluation Authors: Yihang Li, Chenhui Chu Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17260v2 Evaluating meeting effectiveness is crucial for improving organizational productivity. Current approaches rely on post-hoc surveys that yield a single coarse-grained score for an entire meeting. The reliance on manual assessment is inherently limited in scalability, cost, and reproducibility. Moreover, a single score fails to capture the dynamic nature of collaborative discussions. We propose a new paradigm for evaluating meeting effectiveness centered on novel criteria and temporal fine-grained approach. We define effectiveness as the rate of objective achievement over time and assess it for individual topical segments within a meeting. To support this task, we introduce the AMI Meeting Effectiveness (AMI-ME) dataset, a new meta-evaluation dataset containing 2,459 human-annotated segments from 130 AMI Corpus meetings. We also develop an automatic effectiveness evaluation framework that uses a Large Language Model (LLM) as a judge to score each segment's effectiveness relative to the overall meeting objectives. Through substantial experiments, we establish a comprehensive benchmark for this new task and evaluate the framework's generalizability across distinct meeting types, ranging from business scenarios to unstructured discussions. Furthermore, we benchmark end-to-end performance starting from raw speech to measure the capabilities of a complete system. Our results validate the framework's effectiveness and provide strong baselines to facilitate future research in meeting analysis and multi-party dialogue. Our dataset and code will be publicly available. The AMI-ME dataset and the Automatic Evaluation Framework are available at: this URL. --- ### HORIZON: A Benchmark for In-the-wild User Behaviour Modeling Authors: Arnav Goel, Pranjal A Chitale, Bhawna Paliwal, Bishal Santra, Amit Sharma Published: 2026-04-19 Categories: cs.IR, cs.AI, cs.CL arXiv: https://arxiv.org/abs/2604.17259v1 User behavior in the real world is diverse, cross-domain, and spans long time horizons. Existing user modeling benchmarks however remain narrow, focusing mainly on short sessions and next-item prediction within a single domain. Such limitations hinder progress toward robust and generalizable user models. We present HORIZON, a new benchmark that reformulates user modeling along three axes i.e. dataset, task, and evaluation. Built from a large-scale, cross-domain reformulation of Amazon Reviews, HORIZON covers 54M users and 35M items, enabling both pretraining and realistic evaluation of models in heterogeneous environments. Unlike prior benchmarks, it challenges models to generalize across domains, users, and time, moving beyond standard missing-positive prediction in the same domain. We propose new tasks and evaluation setups that better reflect real-world deployment scenarios. These include temporal generalization, sequence-length variation, and modeling unseen users, with metrics designed to assess general user behavior understanding rather than isolated next-item prediction. We benchmark popular sequential recommendation architectures alongside LLM-based baselines that leverage long-term interaction histories. Our results highlight the gap between current methods and the demands of real-world user modeling, while establishing HORIZON as a foundation for research on temporally robust, cross-domain, and general-purpose user models. --- ### REZE: Representation Regularization for Domain-adaptive Text Embedding Pre-finetuning Authors: Seungmin Lee, Jeonghwan Lee, Hyunkuk Lim, Sejoon Kim, Mingi Sung Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17257v2 Recent text embedding models are often adapted to specialized domains via contrastive pre-finetuning (PFT) on a naive collection of scattered, heterogeneous tasks. However, this approach often introduces task-induced bias alongside domain knowledge, leading to uncontrolled representation shifts that distort the pretrained embedding geometry and cause substantial performance degradation. To address this issue, we propose REZE, a representation regularization framework that explicitly controls representation shift during embedding pre-finetuning. REZE operates on the relations of anchor-positive pairs and decomposes them in an eigenspace. It then measures task-wise dispersion along each eigencomponent to identify task-variant directions and applies adaptive soft-shrinkage to suppress task-induced noise while preserving task-invariant semantic structure, without inference-time overhead. Experiments across multiple embedding backbones and specialized benchmarks show that REZE outperforms standard pre-finetuning and isotropy-oriented post-hoc regularization in most settings, remaining stable where existing PFT variants collapse. Embedding space analyses further confirm that REZE induces controlled shifts aligned with the original embedding manifold, underscoring representation shift control as a key principle for robust embedding pre-finetuning under heterogeneous supervision. --- ### A Unified Compliance Aggregator Framework for Automated Multi-Tool Security Assessment of Linux Systems Authors: Sheldon Paul, Izzat Alsmadi Published: 2026-04-19 Categories: cs.CR, cs.LG arXiv: https://arxiv.org/abs/2604.17256v1 Assessing the security posture of modern computing systems typically requires the use of multiple specialized tools. These tools focus on different aspects such as configuration compliance, file integrity, and vulnerability exposure, and their outputs are often difficult to interpret collectively. This paper introduces the Unified Compliance Aggregator (UCA), a framework that integrates several open-source security tools into a single composite score representing overall system security. The proposed framework combines outputs from Lynis, OpenSCAP (STIG and CIS profiles), AIDE, Tripwire, and Nmap NSE. A normalization process converts heterogeneous outputs into a consistent 0 to 100 scale, followed by weighted aggregation. We also introduce a logarithmic scoring model for file integrity measurements to address limitations observed in prior linear approaches. Experiments were conducted on Ubuntu 22.04 across different hardening levels and environments. Results show consistent improvement in composite scores as systems are hardened, while also revealing contrasting behavior between compliance and file integrity tools. Two case studies, a basic web server and a DVWA-based system illustrate how the framework can be applied in practical scenarios. --- ### Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering Authors: Li Zheng, Xin Zhang, Shuyi He, Fei Li, Chong Teng, Jiangming Yang, Donghong Ji, Zhuang Li Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17255v1 Accurate comprehension and controllable generation of emotion and rhetoric are pivotal for enhancing the reasoning capabilities of large language models (LLMs). Existing studies mostly rely on external optimizations, lacking in-depth exploration of internal representation mechanisms, thus failing to achieve fine-grained steering at the neuron level. A handful of works on neurons are confined to emotions, neglecting rhetoric neurons and their intrinsic connections. Traditional neuron masking also exhibits counterintuitive phenomena, making reliable verification of neuron functionality infeasible. To address these issues, we systematically investigate the neurons representation mechanisms and inherent associations of 6 emotion categories and 4 core rhetorical devices. We propose a neuron identification framework that integrates multi-dimensional screening, and design an adaptive masking method incorporating dynamic filtering, attenuation masking, and feedback optimization, enabling reliable causal validation of neuron functionality.Through neuron regulation, we achieve directed induction of non-target sentences and enhancement of emotion tasks via rhetoric neurons. Experiments on 5 commonly used datasets validate the effectiveness of our method, providing a novel paradigm for the fine-grained steering of emotion and rhetoric expressions in LLMs. --- ### Seeing Isn't Believing: Mitigating Belief Inertia via Active Intervention in Embodied Agents Authors: Hanlin Wang, Chak Tou Leong, Jian Wang, Wenjie Li Published: 2026-04-19 Categories: cs.CL, cs.AI, cs.RO arXiv: https://arxiv.org/abs/2604.17252v1 Recent advancements in large language models (LLMs) have enabled agents to tackle complex embodied tasks through environmental interaction. However, these agents still make suboptimal decisions and perform ineffective actions, as they often overlook critical environmental feedback that differs from their internal beliefs. Through a formal probing analysis, we characterize this as belief inertia, a phenomenon where agents stubbornly adhere to prior beliefs despite explicit observations. To address this, we advocate active belief intervention, moving from passive understanding to active management. We introduce the Estimate-Verify-Update (EVU) mechanism, which empowers agents to predict expected outcomes, verify them against observations through explicit reasoning, and actively update prior beliefs based on the verification evidence. EVU is designed as a unified intervention mechanism that generates textual belief states explicitly, and can be integrated into both prompting-based and training-based agent reasoning methods. Extensive experiments across three embodied benchmarks demonstrate that EVU consistently yields substantial gains in task success rates. Further analyses validate that our approach effectively mitigates belief inertia, advancing the development of more robust embodied agents. Our code is available at https://github.com/WangHanLinHenry/EVU. --- ### Bit-Flip Vulnerability of Shared KV-Cache Blocks in LLM Serving Systems Authors: Yuji Yamamoto, Satoshi Matsuura Published: 2026-04-19 Categories: cs.CR, cs.AR, cs.LG arXiv: https://arxiv.org/abs/2604.17249v2 Rowhammer on GPU DRAM has enabled adversarial bit flips in model weights; shared KV-cache blocks in LLM serving systems present an analogous but previously unexamined target. In vLLM's Prefix Caching, these blocks exist as a single physical copy without integrity protection. Using software fault injection under ideal bit targeting, we characterize worst-case severity and identify three properties: (1) Silent divergence - 13 of 16 BF16 bit positions produce coherent but altered outputs, indistinguishable from legitimate responses without a clean baseline. (2) Selective propagation - only requests sharing the targeted prefix are affected. (3) Persistent accumulation - no temporal decay occurs, so cumulative damage grows linearly with subsequent requests. Together, these constitute a threat profile distinct from weight corruption: silent divergence and selective propagation enable detection evasion; persistent accumulation then proceeds unchecked, yielding damage amplification bounded only by how long the block remains cached. A checksum-based countermeasure detects any single-bit corruption at scheduling time, bounding cumulative damage to one batch independent of the block's cache lifetime, with negligible overhead. These results argue for integrity protection of prefix blocks before end-to-end exploitation is demonstrated. --- ### VIBE: Voice-Induced open-ended Bias Evaluation for Large Audio-Language Models via Real-World Speech Authors: Yi-Cheng Lin, Yusuke Hirota, Sung-Feng Huang, Hung-yi Lee Published: 2026-04-19 Categories: eess.AS, cs.CL, cs.SD arXiv: https://arxiv.org/abs/2604.17248v2 Large Audio-Language Models (LALMs) are increasingly integrated into daily applications, yet their generative biases remain underexplored. Existing speech fairness benchmarks rely on synthetic speech and Multiple-Choice Questions (MCQs), both offering a fragmented view of fairness. We propose VIBE, a framework that evaluates generative bias through open-ended tasks such as personalized recommendations, using human-recorded speech. Unlike MCQs, our method allows stereotypical associations to manifest organically without predefined options, making it easily extensible to new tasks. Evaluating 12 state-of-the-art LALMs reveals systematic biases in realistic scenarios. Both gender and accent cues trigger statistically significant distributional shifts, and bias magnitude is strongly task-dependent. --- ### DORA Explorer: Improving the Exploration Ability of LLMs Without Training Authors: Priya Gurjar, Md Farhan Ishmam, Kenneth Marino Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17244v2 Large language model (LLM) agents for sequential decision-making struggle to produce diverse outputs. This leads to insufficient exploration, suboptimal solutions, and repeated actions. Actions are generated at the sequence level, but existing sampling strategies, such as temperature scaling, introduce diversity at the token level, not at the sequence level. We introduce DORA EXPLORER (Diversity-Oriented Ranking of Actions), a training-free, inference-time algorithm for improving exploration in LLM agents. DORA generates multiple candidate actions, scores them using sequence-level log-probability statistics, and samples an action via a tunable exploration parameter. We first study exploration in the classic Multi-Armed Bandit setting, where DORA substantially outperforms temperature-based sampling. Our main evaluation is on the Text Adventure Learning Environment Suite (TALES), where prompting strategies fail to explore but DORA delivers consistent gains across model families, e.g., 31.43% (ReAct) -> 45.5% (DORA) for Qwen-2.5 7B in TextWorld. Beyond exploration, DORA prevents common failures, such as getting stuck in loops. Our project is available at: https://dora-explore.github.io/. --- ### DORA Explorer: Improving the Exploration Ability of LLMs Without Training Authors: Priya Gurjar, Md Farhan Ishmam, Kenneth Marino Published: 2026-04-19 Categories: cs.CL, cs.AI arXiv: https://arxiv.org/abs/2604.17244v1 Despite the rapid progress, LLMs for sequential decision-making (i.e., LLM agents) still struggle to produce diverse outputs. This leads to insufficient exploration, convergence to sub-optimal solutions, and becoming stuck in loops. Such limitations can be problematic in environments that require active exploration to gather information and make decisions. Sampling methods such as temperature scaling introduce token-level randomness but fail to produce enough diversity at the sequence level. We analyze LLM exploration in the classic Multi-Armed Bandit (MAB) setting and the Text Adventure Learning Environment Suite (TALES). We find that current decoding strategies and prompting methods like Chain-of-Thought and Tree-of-Thought are insufficient for robust exploration. To address this, we introduce DORA Explorer (Diversity-Oriented Ranking of Actions), a training-free framework for improving exploration in LLM agents. DORA generates diverse action candidates, scores them using token log-probabilities, and selects actions using a tunable exploration parameter. DORA achieves UCB-competitive performance on MAB and consistent gains across TALES, e.g., improving Qwen2.5-7B's performance from 29.2% to 45.5% in TextWorld. Our project is available at: https://dora-explore.github.io/. --- ### RemoteShield: Enable Robust Multimodal Large Language Models for Earth Observation Authors: Rui Min, Liang Yao, Shiyu Miao, Shengxiang Xu, Yuxuan Liu, Chuanyi Zhang, Shimin Di, Fan Liu Published: 2026-04-19 Categories: cs.CV arXiv: https://arxiv.org/abs/2604.17243v1 A robust Multimodal Large Language Model (MLLM) for Earth Observation should maintain consistent interpretation and reasoning under realistic input variations. However, current Remote Sensing MLLMs fail to meet this requirement. Trained on carefully curated clean datasets, they learn brittle mappings that do not generalize to noisy conditions in operational Earth Observation. Consequently, their performance degrades when confronted with imperfect inputs in deployment. To quantify this vulnerability, we construct a realistic set of multimodal perturbations, including visual degradations such as cloud and fog cover, together with diverse human-centric textual variations ranging from colloquialisms to vague or omitted instructions. Empirical evaluations show that these perturbations significantly impair the visual-semantic reasoning capabilities of leading RS foundation models. To address this limitation, we introduce RemoteShield, a robust Remote Sensing MLLM trained to maintain consistent outputs across realistic input variations. During training, each clean sample is paired with its image-text perturbed variants to form a semantic equivalence cluster. Rather than directly fitting noisy samples, RemoteShield is optimized through preference learning over clean and perturbed conditions within the same cluster. By comparing model responses to clean and corrupted inputs, the model is encouraged to favor stable responses over perturbation-induced failures. This cross-condition alignment helps the model focus on underlying task semantics despite visual degradations and textual noise. Experiments on three Earth Observation tasks show that RemoteShield consistently delivers stronger robustness and cross-condition consistency than representative baselines under realistic multimodal perturbations. --- ### Safe and Policy-Compliant Multi-Agent Orchestration for Enterprise AI Authors: Vinil Pasupuleti, Shyalendar Reddy Allala, Siva Rama Krishna Varma Bayyavarapu, Shrey Tyagi Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17240v1 Enterprise AI systems increasingly deploy multiple intelligent agents across mission-critical workflows that must satisfy hard policy constraints, bounded risk exposure, and comprehensive auditability (SOX, HIPAA, GDPR). Existing coordination methods - cooperative MARL, consensus protocols, and centralized planners - optimize expected reward while treating constraints implicitly. This paper introduces CAMCO (Constraint-Aware Multi-Agent Cognitive Orchestration), a runtime coordination layer that models multi-agent decision-making as a constrained optimization problem. CAMCO integrates three mechanisms: (i) a constraint projection engine enforcing policy-feasible actions via convex projection, (ii) adaptive risk-weighted Lagrangian utility shaping, and (iii) an iterative negotiation protocol with provably bounded convergence. Unlike training-time constrained RL, CAMCO operates as deployment-time middleware compatible with any agent architecture, with policy predicates designed for direct integration with production engines such as OPA. Evaluation across three enterprise scenarios - including comparison against a constrained Lagrangian MARL baseline - demonstrates zero policy violations, risk exposure below threshold (mean ratio 0.71), 92-97% utility retention, and mean convergence in 2.4 iterations. --- ### HeadRank: Decoding-Free Passage Reranking via Preference-Aligned Attention Heads Authors: Juyuan Wang, Chenxing Wang, Yuchen Fang, Huiyun Hu, Junwu Du, Aolin Li, Shunlin Rong, Haijun Wu, Jin Xu, Ligang Liu, Dongliang Liao Published: 2026-04-19 Categories: cs.IR, cs.AI arXiv: https://arxiv.org/abs/2604.17237v2 Decoding-free reranking methods that read relevance signals directly from LLM attention weights offer significant latency advantages over autoregressive approaches, yet suffer from attention score homogenization: middle-context documents receive near-identical scores, destroying the fine-grained distinctions required for ranking. We propose HeadRank, a framework that lifts preference optimization from discrete token space into the continuous attention domain through entropy-regularized head selection, hard adjacent-level preference pairs, and a distribution regularizer that jointly sharpen discriminability in the homogenized middle zone. Depth truncation at the deepest selected layer further reduces inference to $\mathcal{O}(1)$ forward passes. Across 14 benchmarks on three Qwen3 scales (0.6B--4B) using only 211 training queries, HeadRank achieves the highest average NDCG@10 at every scale, outperforming both generative and decoding-free baselines on the majority of benchmarks with 100\% formatting success. At 4B, 57.4\% of relevant middle-zone documents reach the top quartile versus 14.2\% for irrelevant ones -- a 43-percentage-point selectivity gap that demonstrates the effectiveness of attention-space preference alignment for listwise reranking. --- ### Enhancing Zero-shot Personalized Image Aesthetics Assessment with Profile-aware Multimodal LLM Authors: Chun Wang, Chenfeng Wei, Chenyang Liu, Weihong Deng Published: 2026-04-19 Categories: cs.CV, cs.AI arXiv: https://arxiv.org/abs/2604.17233v1 Personalized image aesthetics assessment (PIAA) aims to predict an individual user's subjective rating of an image, which requires modeling user-specific aesthetic preferences. Existing methods rely on historical user ratings for this modeling and therefore struggle when such data are unavailable. We address this zero-shot setting by using user profiles as contextual signals for personalization and adopting a profile-based personalization paradigm. We introduce P-MLLM, a profile-aware multimodal LLM that augments a frozen LLM with selective fusion modules for controlled visual integration. These modules selectively integrate visual information into the model's evolving hidden states during profile-conditioned reasoning, allowing visual information to be incorporated in a profile-aware manner. Experiments on recent PIAA benchmarks show that P-MLLM achieves competitive zero-shot performance and remains effective even with coarse profile information, highlighting the potential of profile-based personalization for zero-shot PIAA. --- ### Fringe Projection Based Vision Pipeline for Autonomous Hard Drive Disassembly Authors: Badrinath Balasubramaniam, Vignesh Suresh, Benjamin Metcalf, Beiwen Li Published: 2026-04-19 Categories: cs.CV, cs.RO arXiv: https://arxiv.org/abs/2604.17231v1 Unrecovered e-waste represents a significant economic loss. Hard disk drives (HDDs) comprise a valuable e-waste stream necessitating robotic disassembly. Automating the disassembly of HDDs requires holistic 3D sensing, scene understanding, and fastener localization, however current methods are fragmented, lack robust 3D sensing, and lack fastener localization. We propose an autonomous vision pipeline which performs 3D sensing using a Fringe Projection Profilometry (FPP) module, with selective triggering of a depth completion module where FPP fails, and integrates this module with a lightweight, real-time instance segmentation network for scene understanding and critical component localization. By utilizing the same FPP camera-projector system for both our depth sensing and component localization modules, our depth maps and derived 3D geometry are inherently pixel-wise aligned with the segmentation masks without registration, providing an advantage over RGB-D perception systems common in industrial sensing. We optimize both our trained depth completion and instance segmentation networks for deployment-oriented inference. The proposed system achieves a box mAP@50 of 0.960 and mask mAP@50 of 0.957 for instance segmentation, while the selected depth completion configuration with the Depth Anything V2 Base backbone achieves an RMSE of 2.317 mm and MAE of 1.836 mm; the Platter Facing learned inference stack achieved a combined latency of 12.86 ms and a throughput of 77.7 Frames Per Second (FPS) on the evaluation workstation. Finally, we adopt a sim-to-real transfer learning approach to augment our physical dataset. The proposed perception pipeline provides both high-fidelity semantic and spatial data which can be valuable for downstream robotic disassembly. The synthetic dataset developed for HDD instance segmentation will be made publicly available. --- ### Yanasse: Finding New Proofs from Deep Vision's Analogies, Part 1 Authors: Alexandre Linhares Published: 2026-04-19 Categories: cs.AI arXiv: https://arxiv.org/abs/2604.17229v1 Project Yanasse presents a method for discovering new proofs of theorems in one area of mathematics by transferring proof strategy patterns (e.g., Lean 4 tactic invocation patterns) from a structurally distant area. The system extracts tactic usage distributions across 27 top-level areas of Mathlib (217,133 proof states), computes z-scores to identify tactics that are heavily used in a source area but rare or absent in a target area, matches source and target proof states via GPU-accelerated NP-hard analogy (running on a MacBook Air via Apple's MPS backend), and then asks an AI reasoning agent to semantically adapt--not symbol-substitute--the source tactics invocation pattern to the target theorem. In this first part of the study, the method is applied to the pair Probability -> Representation Theory, producing 4 Lean-verified new proofs out of 10 attempts (40%). The proofs compile with zero sorry declarations. The key finding is that tactic schemas decompose into a head (domain-gated, rarely transfers) and a modifier (domain-general, often transfers): filter upwards's head fails in representation theory (no Filter structure), but its (LIST) with ω modifier transfers cleanly as ext1 + simp (LIST) + rfl. Crucially, the underlying matching engine--deep vision lib.py--is entirely domain independent: the same optimization code for an NP-hard matching that matches chess positions by analogy matches Lean proof states by analogy, without knowing which domain it is processing. Only a relation extractor is domain-specific. --- ### Revisiting Auxiliary Losses for Conditional Depth Routing: An Empirical Study Authors: Qingwei Lin Published: 2026-04-19 Categories: cs.LG arXiv: https://arxiv.org/abs/2604.17228v1 Conditional depth execution routes a subset of tokens through a lightweight cheap FFN while the remainder execute the standard full FFN at each controlled layer. The central difficulty is gate training: the gate decision must propagate through many layers before it influences the language modeling (LM) loss, so the resulting gradients are weak and noisy. Auxiliary losses are commonly stacked to stabilise training, yet the interactions among them -- particularly between a predictive auxiliary and explicit score supervision -- have not been systematically compared under controlled conditions. We evaluate two gate designs under a 157.5M-parameter decoder-only model with controller-only training, 50% full-path budget, and 3-seed runs on a fineweb-edu subset. The MLP gate (G1) maps the current hidden state to a utility score; the JEPA-guided gate (G3) adds an action-conditional predictor that forecasts, in a low-dimensional latent space, the outcome of executing full vs. cheap per token, aligned against a fixed target head. Under the standard recipe with oracle-style utility regression and pairwise rank supervision (util/rank), G3 improves early-to-mid optimisation over G1 in 3/3 seeds (lower avg LM, faster threshold hits, ~10.3x lower grad norms), with 20k-step endpoint LM within a 0.005 heuristic reference. A key finding (ablation A3): jointly removing util/rank improves best/avg LM and threshold-hit speed in 3/3 seeds for both gates, and the early-to-mid advantage of G3 over G1 disappears. We trace this to an off-policy oracle label that assumes all subsequent layers execute full, whereas gated execution routes only a fraction through full -- making util/rank net-negative under the current recipe. Removing util/rank also cuts the training FLOPs proxy from ~1.53x to ~1.07x full-only (2.87h to 1.75h on a V100-32GB, ~39%). Conclusions are scoped to the studied regime. --- ### A Multi-Agent Approach for Claim Verification from Tabular Data Documents Authors: Rudra Ranajee Saha, Laks V. S. Lakshmanan, Raymond T. Ng Published: 2026-04-19 Categories: cs.CL arXiv: https://arxiv.org/abs/2604.17225v1 We present a novel approach for claim verification from tabular data documents. Recent LLM-based approaches either employ complex pretraining/fine-tuning or decompose verification into subtasks, often lacking comprehensive explanations and generalizability. To address these limitations, we propose a Multi-Agentic framework for Claim verification (MACE) consisting of three specialized agents: Planner, Executor, and Verifier. Instead of elaborate finetuning, each agent employs a zero-shot Chain-of-Thought setup to perform its tasks. MACE produces interpretable verification traces, with the Planner generating explicit reasoning strategies, the Executor providing detailed computation steps, and the Verifier validating the logic. Experiments demonstrate that MACE achieves state-of-the-art (SOTA) performance on two datasets and performs on par with the best models on two others, while achieving 80--100\% of best performance with substantially smaller models: 27--92B parameters versus 235B. This combination of competitive performance, memory efficiency, and transparent reasoning highlights our framework's effectiveness. --- ### LASER: Low-Rank Activation SVD for Efficient Recursion Authors: Ege Çakar, Ketan Ali Raghu, Lia Zheng Published: 2026-04-19 Categories: cs.LG, stat.ML arXiv: https://arxiv.org/abs/2604.17224v1 Recursive architectures such as Tiny Recursive Models (TRMs) perform implicit reasoning through iterative latent computation, yet the geometric structure of these reasoning trajectories remains poorly understood. We investigate the activation manifold of TRMs during recursive unrolling and find that activations occupy an effectively linear, low-dimensional subspace whose principal directions can be tracked dynamically with cheap power iterations. This suggests that weight-sharing concentrates iterative computation along a small number of dominant eigendirections, and we find that this concentration varies sharply across computational sites. We exploit this structure through LASER (Low-Rank Activation SVD for Efficient Recursion), a dynamic compression framework that maintains an evolving low-rank basis via matrix-free subspace tracking with a fidelity-triggered reset mechanism, achieving ${\sim}60\%$ activation memory savings with no statistically significant accuracy degradation. Our analysis raises questions about how recursive architectures allocate representational capacity during implicit reasoning, and whether this concentration can be exploited to improve the efficiency and stability of latent computation. --- ### Region-Affinity Attention for Whole-Slide Breast Cancer Classification in Deep Ultraviolet Imaging Authors: Nagur Shareef Shaik, Teja Krishna Cherukuri, Dong Hye Ye Published: 2026-04-19 Categories: cs.CV, cs.AI, eess.SP arXiv: https://arxiv.org/abs/2604.17222v1 Breast cancer diagnosis demands rapid and precise tools, yet traditional histopathological methods often fall short in intra-operative settings. Deep Ultraviolet (DUV) fluorescence imaging emerges as a transformative approach, offering high-contrast, label-free visualization of whole-slide images (WSIs) with unprecedented detail, surpassing conventional hematoxylin and eosin (H&E) staining in speed and resolution. However, existing deep learning methods for breast cancer classification, predominantly patch-based, fragment spatial context and incur significant preprocessing overhead, limiting their clinical utility. Moreover, standard attention mechanisms, such as Spatial, Squeeze-and-Excitation, Global Context and Guided Context Gating, fail to fully exploit the rich, multi-scale regional relationships inherent in DUV-WSI data, often prioritizing generic feature recalibration over diagnostic specificity. This study introduces a novel Region-Affinity Attention mechanism tailored for DUV-WSI breast cancer classification, processing entire slides without patching to preserve spatial integrity. By modeling local neighbor distances and constructing a full affinity matrix, our method dynamically highlights diagnostically relevant regions, augmented by a contrastive loss to enhance feature discriminability. Evaluated on a dataset of 136 DUV-WSI samples, our approach achieves an accuracy of 92.67 +/- 0.73% and an AUC of 95.97%, outperforming existing attention methods. ---