Papers

New on arXiv

Sat 16 May 2026

264 papers

TILT: Improving Compositional Generation in Diffusion Models with a Model-Intrinsic RewardDiffusion models → Image generation, text-to-image and editing +1
Marking the Wrong Symptoms: Evaluating LLM Watermarks in Medical TextsLLM safety and security → Watermarking generated content and models
When JSON Is Not Enough: Semantic Reliability of Schema-Constrained LLM Ordering AgentsLLM agents and tool use → Tool learning and function calling
VLAMotor: Test-Guided Enhancement of Vision-Language-Action Models via Agent-BasedData SynthesisVision-language-action models → Failure recovery and robustness
Representation Signatures and Risk-Feedback Alignment in LLM Trading AgentsLLM agents and tool use → Financial trading and forecasting agents
MemForest: An Efficient Agent Memory System with Hierarchical Temporal IndexingLLM agents and tool use → Agent memory systems
ObjectCache: Layerwise Object-Storage Retrieval for KV Cache ReuseEfficient LLM inference and serving → KV cache optimization
Provably Learning Diffusion Models under the Manifold Hypothesis: Collapse and RefineDiffusion models → Diffusion foundations and sampling
Prediction Is Not Physics: Learning and Evaluating Conserved Quantities in Neural SimulatorsDiffusion models → Flow matching, time series and science
To Call or Not to Call: Diagnosing Intrinsic Over-Calling Bias in LLM AgentsLLM agents and tool use → Tool learning and function calling +1
Emergence of a Flow-Assisted Casting Strategy for Olfactory Navigation via Memory-Augmented Reinforcement LearningDeep reinforcement learning methods → RL for robot navigation and UAV control
ZeroUnlearn: Few-Shot Knowledge Unlearning in Large Language ModelsLLM safety and security → Machine unlearning
CAREBench: Evaluating LLMs' Emotion Understanding by Assessing Cognitive Appraisal ReasoningLLM reasoning → Audio-language model reasoning
Beyond Execution: Static-Analysis Rewards and Hint-Conditioned Diffusion RL for Code GenerationDiffusion models → Discrete and text diffusion
Why Do Safety Guardrails Degrade Across Languages?LLM safety and security → Safety alignment, refusal and guardrails
TriAxialKV: Toward Extreme Low-Precision KV-Cache Quantization for Agentic Inference TasksLLM agents and tool use → Software engineering coding agents +1
Responsible Agentic AI Requires Explicit ProvenanceLLM agents and tool use → Agent threats, guardrails and governance
STRIDE-AI: A Threat Modeling Framework for Generative AI Security AssessmentLLM safety and security → Prompt injection and agent security
From Imitation to Interaction: Mastering Game of Schnapsen with Shallow Reinforcement LearningDeep reinforcement learning methods → RL for games and combinatorial optimization
MADP: A Multi-Agent Pipeline for Sustainable Document Processing with Human-in-the-LoopLLM agents and tool use → Web, search, and research agents
Sparse Mamba Decoder for Quantum Error Correction: Efficient Defect-Centric Processing of Surface Code SyndromesState-space sequence models (S4, Mamba) → Core SSM architectures for language
Multilingual and Multimodal LLMs in the Wild: Building for Low-Resource LanguagesMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
QPLEX Decision Processes: Formulation via Nonlinear Markov Chains and Optimization via Policy GradientsDeep reinforcement learning methods → RL theory and sample complexity
Weighted Flow Matching and Physics-Informed Nonlinear Filtering for Parameter Estimation in Digital TwinsDiffusion models → Flow matching, time series and science
Contrastive Conceptor Activation Steering (COAST): Unlocking Vision-Language-Action Models through Hidden StatesMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +3
UCSF-PDGM-VQA: Visual Question Answering dataset for brain tumor MRI interpretationMultimodal LLMs / vision-language models → Medical vision-language models
New Wide-Net-Casting Jailbreak Attacks Risk Large ModelsLLM safety and security → Jailbreaking
The Point of No Return: Counterfactual Localization of Deceptive Commitment in Language-Model ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Scientific Logicality Enriched Methodology for LLM Reasoning: A Practice in PhysicsLLM reasoning → Spatial and embodied reasoning
SEMA-RAG: A Self-Evolving Multi-Agent Retrieval-Augmented Generation Framework for Medical ReasoningLLM reasoning → Clinical and biomedical applications +3
HEED: Density-Weighted Residual Alignment for Hybrid Vision-Language Model DistillationMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning +1
ACIL: Auto Chain of Thoughts for In-Context LearningLLM reasoning → Prompting and chain-of-thought reasoning
The Learnability Gap in Medical Latent DiffusionDiffusion models → Medical imaging and segmentation
Taming Audio VAEs via Target-KL RegularizationDiffusion models → Audio, speech and music
Can LLMs Think Like Consumers? Benchmarking Crowd-Level Reaction Reconstruction with ConsumerSimBenchLLM reasoning → Reasoning benchmarks and evaluation
S-Bus: Automatic Read-Set Reconstruction for Multi-Agent LLM State CoordinationLLM agents and tool use → Agent memory systems
A Red Teaming Framework for Evaluating Robustness of AI-enabled Security Orchestration, Automation, and Response SystemsLLM safety and security → Prompt injection and agent security
RAGA: Reading-And-Graph-building-Agent for Autonomous Knowledge Graph Construction and Retrieval-Augmented GenerationRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG +1
AnchorDiff: Topology-Aware Masked Diffusion with Confidence-based Rewriting for Radiology Report GenerationDiffusion models → Medical imaging and segmentation
EPIC-Bench: A Perception-Centric Benchmark for Fine-Grained Embodied Visual Grounding in Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
PyraVid: Hierarchical Multimodal Memory for Long-Horizon Video ReasoningLLM agents and tool use → Multimodal and embodied agents
Towards Human-Level Book-Writing CapabilityLLM agents and tool use → Multimodal and embodied agents
Learning Multi-Timescale Abstractions for Hierarchical Combinatorial PlanningWorld models → Latent predictive world models +1
1GC-7RC: One Graphic Card -- Seven Research Challenges! How Good Are AI Agents at Doing Your Job?LLM agents and tool use → Web, search, and research agents
Agentic AI Translate: An Agentic Translator Prototype for Translation as Communication DesignLLM agents and tool use → General LLM agent frameworks and benchmarks
D2Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning +1
Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented GenerationRetrieval-augmented generation → Privacy and security of RAG +1
Why Do Reasoning Models Lose Coverage? The Role of Data and Forks in the RoadLLM reasoning → Test-time compute and efficient thinking
When Dynamics Shift, Robust Task Inference Wins: Offline Imitation Learning with Behavior Foundation Models RevisitedOffline RL and imitation learning from data → Imitation learning methods
Learning-Zone Energy: Online Data Selection for Efficient RL Post-TrainingLLM reasoning → Reinforcement learning for reasoning
Ranking-Aware Calibration for Reliable Multimodal Reinforcement LearningLLM reasoning → Multimodal reasoning +2
DreamEdit3D: Personalization of Multi-View Diffusion Models for 3D EditingDiffusion models → 3D, point cloud and scene generation
Skills on the Fly: Test-Time Adaptive Skill Synthesis for LLM AgentsLLM reasoning → Reasoning for agents, planning and code +1
Securing LLM Agents Need Intent-to-Execution IntegrityLLM agents and tool use → Agent threats, guardrails and governance
OmniVL-Guard Pro: A Tool-Augmented Agent for Omnibus Vision-Language ForensicsLLM agents and tool use → Multimodal and embodied agents
Beyond Point-Wise Matching: Structural Representation Alignment for Accelerating Diffusion TransformersDiffusion models → Diffusion foundations and sampling
Effort as Ceiling, Not Dial: Reasoning Budget Does Not Modulate Cognitive Cost Alignment Between Humans and Large Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training StepsEfficient LLM inference and serving → KV cache optimization
HighSync: High-Quality Lip Synchronization via Latent Diffusion ModelsDiffusion models → Audio, speech and music
LASAR: Towards Spatio-temporal Reasoning with Latent Cognitive MapLLM reasoning → Spatial and embodied reasoning
The Alpha Illusion: Reported Alpha from LLM Trading Agents Should Not Be Treated as Deployment EvidenceLLM agents and tool use → Financial trading and forecasting agents
Beyond Safety Filtering: Control Barrier Function-Informed Reinforcement Learning for Connected and Automated VehiclesDeep reinforcement learning methods → RL for autonomous driving and traffic
DriveSafe: A Framework for Risk Detection and Safety Suggestions in Driving ScenariosEnd-to-end autonomous driving → VLMs for autonomous driving
SE-GA: Memory-Augmented Self-Evolution for GUI AgentsLLM agents and tool use → GUI and mobile agents
Towards Generalized Image Manipulation Localization via Score-based ModelDiffusion models → Image generation, text-to-image and editing
Reasoning Can Be Restored by Correcting a Few Decision TokensLLM reasoning → Test-time compute and efficient thinking
HAD: Hallucination-Aware Diffusion Priors for 3D ReconstructionDiffusion models → Medical imaging and segmentation
SADP: Subgoal-Aware Diffusion Policy for Long-Horizon Manipulation Learned from Foundation Model Generated DemonstrationsDiffusion models → Diffusion and flow policies for robotics
MixSD: Mixed Contextual Self-Distillation for Knowledge InjectionKnowledge distillation → Knowledge distillation for LLMs
Plan First, Diffuse Later: Extrinsic Graph Guidance for Long-Horizon Diffusion PlanningDiffusion models → Diffusion and flow policies for robotics
Prefix-Adaptive Block Diffusion for Efficient Document RecognitionDiffusion models → Discrete and text diffusion +1
Pedestrian-Aware LLM-Driven Behavioral Planning for Autonomous VehiclesEnd-to-end autonomous driving → Large language models for driving
Thinking with Patterns: Breaking the Perceptual Bottleneck in Visual Planning via Pattern InductionWorld models → Latent predictive world models +2
Sketch Then Paint: Hierarchical Reinforcement Learning for Diffusion Multi-Modal Large Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Constrained Code Generation with Discrete DiffusionDiffusion models → Discrete and text diffusion
Decoupling KL and Trajectories: A Unified Perspective for SFT, DAgger, Offline RL, and OPD in LLM DistillationOffline RL and imitation learning from data → Offline RL algorithms +1
Confidence Geometry Reveals Trace-Level Correctness in Large Language Model ReasoningLLM reasoning → Test-time compute and efficient thinking
Multi-Paradigm Agent Interaction in Practice:A Systematic Analysis of Generator-Evaluator, ReAct Loop,and Adversarial Evaluation in the buddyMe FrameworkLLM agents and tool use → General LLM agent frameworks and benchmarks
AgentKernelArena: Generalization-Aware Benchmarking of GPU Kernel Optimization AgentsLLM agents and tool use → Software engineering coding agents
TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool CompositionLLM agents and tool use → Tool learning and function calling
Accelerating Rectified Flow Models via Trajectory-Aware CachingDiffusion models → Diffusion foundations and sampling
The Unlearnability Phenomenon in RLVR for Language ModelsLLM reasoning → Reinforcement learning for reasoning
Lever: Speculative LLM Inference on SmartphonesEfficient LLM inference and serving → Speculative decoding
Distinguishable Deletion: Unifying Knowledge Erasure and Refusal for Large Language Model UnlearningLLM safety and security → Machine unlearning
Axial-Relation Guided Fusion State Space Model for Optical-Elevation Sensing Image SegmentationState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
NeuroMAS: Multi-Agent Systems as Neural Networks with Joint Reinforcement LearningLLM agents and tool use → General LLM agent frameworks and benchmarks
Learning Unbiased Permutations via Flow MatchingDiffusion models → Flow matching, time series and science
State Contamination in Memory-Augmented LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance
EVA01: Unified Native 3D Understanding and Generation via Mixture-of-TransformersMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
DriveSafer: End-to-End Autonomous Driving with Safety GuidanceEnd-to-end autonomous driving → Adversarial scenarios and safety testing
CAB: Accelerating Flow and Diffusion Sampling via Rectification and Corrected Adams-BashforthDiffusion models → Diffusion foundations and sampling
DiRotQ: Rotation-Aware Quantization for 4-bit Diffusion TransformersDiffusion models → Diffusion foundations and sampling
PopuLoRA: Co-Evolving LLM Populations for Reasoning Self-PlayLLM reasoning → Reinforcement learning for reasoning +1
Baba in Wonderland: Online Self-Supervised Dynamics Discovery for Executable World ModelsWorld models → World models for language agents
Reverse Item Response Theory for Sparsity-Robust Ranking in Fragmented Cancer Drug-Response Matrices
Reliable LLM-Powered Decision Engines for Large-Scale Supply Chain Operations: Architecture, Safety, and Performance Guarantees
Evaluating the Impact of Reviewer Guideline Design on LLM-Based Automated Peer Review
Spectral Flow Certificates for Depth-Aware Long-Range Propagation in Graph Neural Networks
Position: AI/ML Deepfake Research is Misaligned with AI-Generated Non-Consensual Intimate Imagery (AIG-NCII)
ProbSPARQL: Querying Knowledge Graphs with Multi-dimensional, Uncertain Numeric Data

All 264