Papers

New on arXiv

Sat 4 Apr 2026

182 papers

Parallel LLM Reasoning for Bias-Resilient, Robust Conceptual AbstractionLLM reasoning → Reasoning benchmarks and evaluation
Differences in Text Generated by Diffusion and Autoregressive Language ModelsDiffusion models → Discrete and text diffusion
From Euler to Dormand-Prince: ODE Solvers for Flow Matching Generative ModelsDiffusion models → Flow matching, time series and science
Safety, Security, and Cognitive Risks in State-Space Models: A Systematic Threat Analysis with Spectral, Stateful, and Capacity AttacksState-space sequence models (S4, Mamba) → Core SSM architectures for language
ML-based approach to classification and generation of structured light propagation in turbulent mediaDiffusion models → Inverse problems and image restoration
Design Conditions for Intra-Group Learning of Sequence-Level Rewards: Token Gradient CancellationLLM reasoning → Reinforcement learning for reasoning
CURE:Circuit-Aware Unlearning for LLM-based RecommendationLLM safety and security → Membership inference attacks +1
Squeez: Task-Conditioned Tool-Output Pruning for Coding AgentsLLM agents and tool use → Software engineering coding agents
Measuring the Permission Gate: A Stress-Test Evaluation of Claude Code's Auto ModeLLM agents and tool use → Agent threats, guardrails and governance
RCP: Representation Consistency Pruner for Mitigating Distribution Shift in Large Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
MG2-RAG: Multi-Granularity Graph for Multimodal Retrieval-Augmented GenerationMultimodal LLMs / vision-language models → Document understanding and multimodal RAG +1
LLM-Agent-based Social Simulation for Attitude DiffusionLLM agents and tool use → Social simulation and multi-agent debate
FeynmanBench: Benchmarking Multimodal LLMs on Diagrammatic Physics ReasoningLLM reasoning → Spatial and embodied reasoning +1
Provable Multi-Task Reinforcement Learning: A Representation Learning Framework with Low Rank RewardsDeep reinforcement learning methods → Deep RL algorithms and representations
PolySwarm: A Multi-Agent Large Language Model Framework for Prediction Market Trading and Latency ArbitrageLLM agents and tool use → Financial trading and forecasting agents
PhaseFlow4D: Physically Constrained 4D Beam Reconstruction via Feedback-Guided Latent DiffusionDiffusion models → Flow matching, time series and science
When Models Know More Than They Say: Probing Analogical Reasoning in LLMsLLM reasoning → Reasoning benchmarks and evaluation
SODA: Semi On-Policy Black-Box Distillation for Large Language ModelsKnowledge distillation → Knowledge distillation for LLMs
Your Agent is More Brittle Than You Think: Uncovering Indirect Injection Vulnerabilities in Agentic LLMsLLM agents and tool use → Agent threats, guardrails and governance
Where to Steer: Input-Dependent Layer Selection for Steering Improves LLM AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods +1
Training a Student Expert via Semi-Supervised Foundation Model DistillationKnowledge distillation → Detection and dense prediction distillation
SPARK-IL: Spectral Retrieval-Augmented RAG for Knowledge-driven Deepfake Detection via Incremental LearningRetrieval-augmented generation → Multimodal RAG
Representational Collapse in Multi-Agent LLM Committees: Measurement and Diversity-Aware ConsensusLLM agents and tool use → Social simulation and multi-agent debate
When AI Agents Disagree Like Humans: Reasoning Trace Analysis for Human-AI Collaborative ModerationLLM reasoning → Reasoning for agents, planning and code +1
Automated Conjecture Resolution with Formal VerificationLLM reasoning → Mathematical reasoning
CountsDiff: A Diffusion Model on the Natural Numbers for Generation and Imputation of Count-Based DataDiffusion models → Flow matching, time series and science
When Does Multimodal AI Help? Diagnostic Complementarity of Vision-Language Models and CNNs for Spectrum Management in Satellite-Terrestrial NetworksLLM reasoning → Multimodal reasoning +1
RL-Driven Sustainable Land-Use Allocation for the Lake Malawi BasinDeep reinforcement learning methods → RL for energy, networks and operations
ITIScore: An Image-to-Text-to-Image Rating Framework for the Image Captioning Ability of MLLMsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Build on Priors: Vision--Language--Guided Neuro-Symbolic Imitation Learning for Data-Efficient Real-World Robot ManipulationOffline RL and imitation learning from data → Robot manipulation imitation learning
VidNum: Diagnosing VLM Failure Modes in Video-Grounded Numerical ReasoningLLM reasoning → Multimodal reasoning +1
FunFact: Building Probabilistic Functional 3D Scene Graphs via Factor-Graph ReasoningLLM reasoning → Spatial and embodied reasoning
LightThinker++: From Reasoning Compression to Memory ManagementLLM reasoning → Test-time compute and efficient thinking
Unlocking Prompt Infilling Capability for Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Are LLM-Based Retrievers Worth Their Cost? An Empirical Study of Efficiency, Robustness, and Reasoning OverheadLLM reasoning → Knowledge- and tool-augmented reasoning
OASES: Outcome-Aligned Search-Evaluation Co-Training for Agentic SearchLLM agents and tool use → Web, search, and research agents
DiffSparse: Accelerating Diffusion Transformers with Learned Token SparsityDiffusion models → Diffusion foundations and sampling
User Simulator-Guided Multi-Turn Preference Optimization for Reasoning LLM-based Conversational RecommendationLLM reasoning → LLM reasoning for recommendation
MMP-Refer: Multimodal Path Retrieval-augmented LLMs For Explainable RecommendationRetrieval-augmented generation → RAG for recommendation
Document-Level Numerical Reasoning across Single and Multiple Tables in Financial ReportsLLM reasoning → Reasoning for agents, planning and code
TableVision: A Large-Scale Benchmark for Spatially Grounded Reasoning over Complex Hierarchical TablesLLM reasoning → Spatial and embodied reasoning
Beyond Retrieval: Modeling Confidence Decay and Deterministic Agentic Platforms in Generative Engine OptimizationLLM agents and tool use → Web, search, and research agents
Joint Behavior-guided and Modality-coherence Conditional Graph Diffusion Denoising for Multi Modal RecommendationDiffusion models → Diffusion recommenders
CAGMamba: Context-Aware Gated Cross-Modal Mamba Network for Multimodal Sentiment AnalysisState-space sequence models (S4, Mamba) → Speech and audio SSMs
Stabilizing Unsupervised Self-Evolution of MLLMs via Continuous Softened Retracing reSamplingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Delayed homomorphic reinforcement learning for environments with delayed feedbackDeep reinforcement learning methods → Deep RL algorithms and representations
Single-agent vs. Multi-agents for Automated Video Analysis of On-Screen Collaborative Learning BehaviorsLLM agents and tool use → Multimodal and embodied agents
Human-Robot Copilot for Data-Efficient Imitation LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
Multi-Robot Multi-Queue Control via Exhaustive Assignment Actor-Critic LearningDeep reinforcement learning methods → RL for games and combinatorial optimization
Stochastic Generative Plug-and-Play PriorsDiffusion models → Inverse problems and image restoration
AttackEval: A Systematic Empirical Study of Prompt Injection Attack Effectiveness Against Large Language ModelsLLM safety and security → Prompt injection and agent security
Unveiling Language Routing Isolation in Multilingual MoE Models for Interpretable Subnetwork AdaptationMixture-of-experts → Expert routing strategies
SecPI: Secure Code Generation with Reasoning Models via Security Reasoning InternalizationLLM reasoning → Reasoning for agents, planning and code
MultiPress: A Multi-Agent Framework for Interpretable Multimodal News ClassificationRetrieval-augmented generation → Agentic RAG, agents and memory
HAD: Combining Hierarchical Diffusion with Metric-Decoupled RL for End-to-End DrivingEnd-to-end autonomous driving → Imitation-learning end-to-end planners
Selective Forgetting for Large Reasoning ModelsLLM reasoning → Reasoning for agents, planning and code +1
When Do Hallucinations Arise? A Graph Perspective on the Evolution of Path Reuse and Path CompressionLLM reasoning → Hallucination detection and mitigation
Focus Matters: Phase-Aware Suppression for Hallucination in Vision-Language ModelsLLM reasoning → Hallucination detection and mitigation +1
CRAFT: Video Diffusion for Bimanual Robot Data GenerationDiffusion models → Diffusion and flow policies for robotics +1
AgenticFlict: A Large-Scale Dataset of Merge Conflicts in AI Coding Agent Pull Requests on GitHubLLM agents and tool use → Software engineering coding agents
Drift-Based Policy Optimization: Native One-Step Policy Learning for Online Robot ControlDiffusion models → Diffusion and flow policies for robotics +1
Rethinking Token Prediction: Tree-Structured Diffusion Language ModelLLM reasoning → Diffusion reasoning and test-time scaling +1
LangFIR: Discovering Sparse Language-Specific Features from Monolingual Data for Language SteeringMechanistic interpretability → SAE features and their applications
Explainable Model Routing for Agentic WorkflowsLLM agents and tool use → General LLM agent frameworks and benchmarks
Optimizing Neurorobot Policy under Limited Demonstration Data through Preference RegretOffline RL and imitation learning from data → Robot learning from demonstration
Pseudo-Siamese Network for Planning in Target-Oriented Proactive Dialogues
Improving Quantized Model Performance in Qualitative Analysis with Multi-Pass Prompt Verification
Leveraging Large Language Models for Sentiment Analysis: Multi-Modal Analysis of Decentraland's MANA Token
Adaptive Alarm Threshold Prediction in 4G Mobile Networks: A Percentile-Guided Deep Learning Framework with Interpretable Outputs
Fast Log-Domain Sinkhorn Optimal Transport with Warp-Level GPU Reductions
Sparse Regression under Correlation and Weak Signals: A Reproducible Benchmark of Classical and Bayesian Methods
Polynomial-Time Optimal Group Selection via the Double-Commutator Eigenvalue Problem
GCA-BULF: A Bottom-Up Framework for Short-Term Load Forecasting Using Grouped Critical Appliances
Interpretable Fuzzy Modeling Reveals Population-Level Representation Differences in P300 Brain Computer Interfaces Across Neurodivergent and Neurotypical Cohorts
AutoCompress: Critical Layer Isolation for Efficient Transformer Compression
Functional Similarity Metric for Neural Networks: Overcoming Parametric Ambiguity via Activation Region Analysis
Method for Aggregating Unstructured Data Using Large Language Models
Aletheia: Gradient-Guided Layer Selection for Efficient LoRA Fine-Tuning Across Architectures
Portfolio Optimization Proxies under Label Scarcity and Regime Shifts via Bayesian and Deterministic Students under Semi-Supervised Sandwich Training
Adaptive Rigor in AI System Evaluation using Temperature-Controlled Verdict Aggregation via Generalized Power Mean
Negotiating Privacy with Smart Voice Assistants: Risk-Benefit and Control-Acceptance Tensions
EDGE-Shield: Efficient Denoising-staGE Shield for Violative Content Filtering via Scalable Reference-Based Matching
Territory Paint Wars: Diagnosing and Mitigating Failure Modes in Competitive Multi-Agent PPO
An Imbalanced Dataset with Multiple Feature Representations for Studying Quality Control of Next-Generation Sequencing
Towards Predicting Multi-Vulnerability Attack Chains in Software Supply Chains from Software Bill of Materials Graphs
StrADiff: A Structured Source-Wise Adaptive Diffusion Framework for Linear and Nonlinear Blind Source Separation
A Theory-guided Weighted L2 Loss for solving the BGK model via Physics-informed neural networks
Belief Dynamics for Detecting Behavioral Shifts in Safe Collaborative Manipulation
Latency-Aware Resource Allocation over Heterogeneous Networks: A Lorentz-Invariant Market Mechanism
Lotka-Sharpe Neural Operators for Control of Population PDEs
Regime-Calibrated Fleet Repositioning with a Spatial Queue-Regret Decomposition
LLM-generated personalized nudges for improving pro-environmental behavior: Field evidence from resource conservation
Learning 3D Reconstruction with Priors in Test Time
Neural Processes Maintain Calibrated Biomass Estimates Across Spatiotemporal Gaps and Disturbance
SecureAFL: Secure Asynchronous Federated Learning
A Bayesian Information-Theoretic Approach to Data Attribution
Understanding When Poisson Log-Normal Models Outperform Penalized Poisson Regression for Microbiome Count Data
Collapse-Free Prototype Readout Layer for Transformer Encoders
Explainability-Guided Adversarial Attacks on Transformer-Based Malware Detectors Using Control Flow Graphs
New insights into Elo algorithm for practitioners and statisticians

All 182