Papers

New on arXiv

Fri 27 Mar 2026

377 papers

Clinical Reasoning AI for Oncology Treatment Planning: A Multi-Specialty Case-Based EvaluationLLM reasoning → Clinical and biomedical applications
From Actions to Understanding: Conformal Interpretability of Temporal Concepts in LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks +1
Cognis: Context-Aware Memory for Conversational AI AgentsLLM agents and tool use → Agent memory systems
TTKV: Temporal-Tiered KV Cache for Long-Context LLM InferenceEfficient LLM inference and serving → KV cache optimization
Accelerating PayPal's Commerce Agent with Speculative Decoding: An Empirical Study on EAGLE3 with Fine-Tuned Nemotron ModelsLLM agents and tool use → Software engineering coding agents +1
OThink-SRR1: Search, Refine and Reasoning with Reinforced Learning for Large Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Two-dimensional early exit optimisation of LLM inferenceEfficient LLM inference and serving → Early exit and edge inference
Disentangled Robot Learning via Separate Forward and Inverse Dynamics PretrainingVision-language-action models → Human-video and latent-action pretraining
StressWeb: A Diagnostic Benchmark for Web Agent Robustness under Realistic Interaction VariabilityLLM agents and tool use → Web, search, and research agents
Distributionally Robust Token Optimization in RLHFLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
GAN-Enhanced Deep Reinforcement Learning for Semantic-Aware Resource Allocation in 6G Network SlicingDeep reinforcement learning methods → RL for energy, networks and operations
Scaling Teams or Scaling Time? Memory Enabled Lifelong Learning in LLM Multi-Agent SystemsLLM agents and tool use → Agent memory systems
Trustworthy AI-Driven Dynamic Hybrid RIS: Joint Optimization and Reward Poisoning-Resilient Control in Cognitive MISO NetworksDeep reinforcement learning methods → RL for energy, networks and operations
DarwinNet: An Evolutionary Network Architecture for Agent-Driven Protocol SynthesisLLM agents and tool use → General LLM agent frameworks and benchmarks
Generalizable Foundation Models for Calorimetry via Mixtures-of-Experts and Parameter Efficient Fine TuningMixture-of-experts → MoE theory and gating functions
Not Search, But Scan: Benchmarking MLLMs on Scan-Oriented Academic Paper ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning +1
Unsupervised Behavioral Compression: Learning Low-Dimensional Policy Manifolds through State-Occupancy MatchingDeep reinforcement learning methods → Deep RL algorithms and representations
Unified Number-Free Text-to-Motion Generation Via Flow MatchingDiffusion models → Video generation
TAPS: Task Aware Proposal Distributions for Speculative SamplingEfficient LLM inference and serving → Speculative decoding
GUIDED: Granular Understanding via Identification, Detection, and Discrimination for Fine-Grained Open-Vocabulary Object DetectionMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection
RASPRef: Retrieval-Augmented Self-Supervised Prompt Refinement for Large Reasoning ModelsLLM reasoning → Prompting and chain-of-thought reasoning +1
Beyond Freshness and Semantics: A Coupon-Collector Framework for Effective Status UpdatesDeep reinforcement learning methods → RL for energy, networks and operations
ROSClaw: An OpenClaw ROS 2 Framework for Agentic Robot Control and InteractionLLM agents and tool use → Multimodal and embodied agents
On the Reliability Limits of LLM-Based Multi-Agent PlanningLLM agents and tool use → General LLM agent frameworks and benchmarks
A Provable Energy-Guided Test-Time Defense Boosting Adversarial Robustness of Large Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Online Statistical Inference of Constant Sample-averaged Q-LearningDeep reinforcement learning methods → RL theory and sample complexity
ASTER -- Agentic Science Toolkit for Exoplanet ResearchLLM agents and tool use → Web, search, and research agents
The Observability Gap: Why Output-Level Human Feedback Fails for LLM Coding AgentsLLM agents and tool use → Multimodal and embodied agents +1
FusionAgent: A Multimodal Agent with Dynamic Model Selection for Human RecognitionLLM agents and tool use → Multimodal and embodied agents
Stable Reasoning, Unstable Responses: Mitigating LLM Deception via Stability AsymmetryLLM reasoning → Prompting and chain-of-thought reasoning
GISclaw: A Comprehensive Open-Source LLM Agent System for Realistic Multi-Step Geospatial AnalysisLLM agents and tool use → Multimodal and embodied agents
SpatialAnt: Autonomous Zero-Shot Robot Navigation via Active Scene Reconstruction and Visual AnticipationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Hybrid Diffusion Model for Breast Ultrasound Image AugmentationDiffusion models → Medical imaging and segmentation
VLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy DistillationVision-language-action models → Reinforcement learning and post-training
Learning to Commit: Generating Organic Pull Requests via Online Repository MemoryLLM agents and tool use → Software engineering coding agents
GaussianGPT: Towards Autoregressive 3D Gaussian Scene GenerationDiffusion models → 3D, point cloud and scene generation
PerceptionComp: A Video Benchmark for Complex Perception-Centric ReasoningLLM reasoning → Multimodal reasoning
Vision2Web: A Hierarchical Benchmark for Visual Website Development with Agent VerificationLLM agents and tool use → Web, search, and research agents
Make Geometry Matter for Spatial ReasoningLLM reasoning → Spatial and embodied reasoning +1
Deception and Communication in Autonomous Multi-Agent Systems: An Experimental Study with Among UsLLM agents and tool use → Social simulation and multi-agent debate
Meta-Adaptive Beam Search Planning for Transformer-Based Reinforcement Learning Control of UAVs with Overhead Manipulators under Flight DisturbancesDeep reinforcement learning methods → RL for robot navigation and UAV control
VGGRPO: Towards World-Consistent Video Generation with 4D Latent RewardDiffusion models → 3D, point cloud and scene generation
The Limits of Learning from Pictures and Text: Vision-Language Models and Embodied Scene UnderstandingMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
From Synthetic Data to Real Restorations: Diffusion Model for Patient-specific Dental Crown CompletionDiffusion models → Medical imaging and segmentation
MA-Bench: Towards Fine-grained Micro-Action UnderstandingMultimodal LLMs / vision-language models → Multimodal emotion understanding
GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified FlowDiffusion models → Diffusion for compression and wireless
Machine Unlearning under Retain-Forget EntanglementLLM safety and security → Machine unlearning
SemConFlow: Semantic Grounding of Holistic Co-Speech Gesture Generation with Contrastive Flow-MatchingDiffusion models → Audio, speech and music
A Lyapunov Analysis of Softmax Policy Gradient for Stochastic BanditsDeep reinforcement learning methods → RL theory and sample complexity
PAPO: Stabilizing Rubric Integration Training via Decoupled Advantage NormalizationLLM reasoning → Reinforcement learning for reasoning +1
CADSmith: Multi-Agent CAD Generation with Programmatic Geometric ValidationLLM agents and tool use → Multimodal and embodied agents
AIRA_2: Overcoming Bottlenecks in AI Research AgentsLLM agents and tool use → Web, search, and research agents
TCM-Serve: Modality-aware Scheduling for Multimodal Large Language Model InferenceMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Entanglement as Memory: Mechanistic Interpretability of Quantum Language ModelsMechanistic interpretability → Circuits and mechanistic analysis
ClipTTT: CLIP-Guided Test-Time Training Helps LVLMs See BetterMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Addressing Ambiguity in Imitation Learning through Product of Experts based Negative FeedbackOffline RL and imitation learning from data → Imitation learning methods
Automatic feature identification in least-squares policy iteration using the Koopman operator frameworkDeep reinforcement learning methods → RL theory and sample complexity
120 Minutes and a Laptop: Minimalist Image-goal Navigation via Unsupervised Exploration and Offline RLOffline RL and imitation learning from data → Language-conditioned policies and VLAs
Reconstructing Quantum Dot Charge Stability Diagrams with Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Why Models Know But Don't Say: Chain-of-Thought Faithfulness Divergence Between Thinking Tokens and Answers in Open-Weight Reasoning ModelsLLM reasoning → Prompting and chain-of-thought reasoning
WiMamba: Linear-Scale Wireless Foundation ModelState-space sequence models (S4, Mamba) → Core SSM architectures for language
Dynamic Token Compression for Efficient Video Understanding through Reinforcement LearningMultimodal LLMs / vision-language models → Video understanding with MLLMs +2
LLaDA-TTS: Unifying Speech Synthesis and Zero-Shot Editing via Masked Diffusion ModelingDiffusion models → Audio, speech and music
HandVQA: Diagnosing and Improving Fine-Grained Spatial Reasoning about Hands in Vision-Language ModelsLLM reasoning → Spatial and embodied reasoning +1
MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion ModelDiffusion models → Diffusion foundations and sampling
Reflect to Inform: Boosting Multimodal Reasoning via Information-Gain-Driven VerificationLLM reasoning → Multimodal reasoning +1
CALRK-Bench: Evaluating Context-Aware Legal Reasoning in Korean LawLLM reasoning → Reasoning for agents, planning and code
Mitigating the Reasoning Tax in Vision-Language Fine-Tuning with Input-Adaptive Depth AggregationLLM reasoning → Multimodal reasoning +1
From Human Cognition to Neural Activations: Probing the Computational Primitives of Spatial Reasoning in LLMsLLM reasoning → Spatial and embodied reasoning
DFM-VLA: Iterative Action Refinement for Robot Manipulation via Discrete Flow MatchingVision-language-action models → Efficient VLA inference and adaptation +1
SALMUBench: A Benchmark for Sensitive Association-Level Multimodal UnlearningLLM safety and security → Machine unlearning
Knowdit: Agentic Smart Contract Vulnerability Detection with Auditing Knowledge SummarizationLLM agents and tool use → Agent threats, guardrails and governance
GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play AnnotationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +2
Topology-Aware Graph Reinforcement Learning for Energy Storage Systems Optimal Dispatch in Distribution NetworksDeep reinforcement learning methods → RL for energy, networks and operations
DRUM: Diffusion-based Raydrop-aware Unpaired Mapping for Sim2Real LiDAR SegmentationDiffusion models → 3D, point cloud and scene generation
Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management SystemsOffline RL and imitation learning from data → Decision transformers and sequence modeling +2
A Universal Vibe? Finding and Controlling Language-Agnostic Informal Register with SAEsMechanistic interpretability → SAE features and their applications
Ask or Assume? Uncertainty-Aware Clarification-Seeking in Coding AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Clawed and Dangerous: Can We Trust Open Agentic Systems?LLM agents and tool use → Agent threats, guardrails and governance
Towards GUI Agents: Vision-Language Diffusion Models for GUI GroundingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +2
ClinicalAgents: Multi-Agent Orchestration for Clinical Decision Making with Dual-MemoryLLM agents and tool use → Medical and scientific domain agents
Can AI Scientist Agents Learn from Lab-in-the-Loop Feedback? Evidence from Iterative Perturbation DiscoveryLLM agents and tool use → Web, search, and research agents
Gaussian Shannon: High-Precision Diffusion Model Watermarking Based on CommunicationDiffusion models → Robustness, privacy and adversarial aspects +1
Efficient Few-Shot Learning for Edge AI via Knowledge Distillation on MobileViTKnowledge distillation → General knowledge distillation methods
Simulating Novice Students Using Machine Unlearning and Relearning in Large Language ModelsLLM safety and security → Machine unlearning
Beyond Where to Look: Trajectory-Guided Reinforcement Learning for Multimodal RLVRLLM reasoning → Multimodal reasoning +1
SkinGPT-X: A Self-Evolving Collaborative Multi-Agent System for Transparent and Trustworthy Dermatological DiagnosisLLM agents and tool use → Medical and scientific domain agents +1
Rethinking Recommendation Paradigms: From Pipelines to Agentic Recommender SystemsLLM reasoning → LLM reasoning for recommendation +1
Dynamic Tokenization via Reinforcement Patching: End-to-end Training and Zero-shot TransferDeep reinforcement learning methods → RL training of language and agent models
Learnable Instance Attention Filtering for Adaptive Detector DistillationKnowledge distillation → Detection and dense prediction distillation
AgenticRS-Architecture: System Design for Agentic Recommender SystemsLLM agents and tool use → Agentic recommender systems
When Identities Collapse: A Stress-Test Benchmark for Multi-Subject PersonalizationDiffusion models → Image generation, text-to-image and editing
Not All Entities are Created Equal: A Dynamic Anonymization Framework for Privacy-Preserving RAGRetrieval-augmented generation → Privacy and security of RAG
Seeing Like Radiologists: Context- and Gaze-Guided Vision-Language Pretraining for Chest X-raysLLM reasoning → Clinical and biomedical applications +1
Retrieval-Augmented Generation Based Nurse Observation ExtractionRetrieval-augmented generation → Medical and biomedical RAG
Where and How to Prune: An Empirical Study of Visual Token Pruning for GUI Agent NavigationLLM agents and tool use → GUI and mobile agents +1
AgentCollab: A Self-Evaluation-Driven Collaboration Paradigm for Efficient LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Knowledge is Power: Advancing Few-shot Action Recognition with Multimodal Semantics from MLLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs
VLAgeBench: Benchmarking Large Vision-Language Models for Zero-Shot Human Age EstimationMultimodal LLMs / vision-language models → Social bias and fairness in VLMs
Cone-Beam CT Image Quality Enhancement Using A Latent Diffusion Model Trained with Simulated CBCT ArtifactsDiffusion models → Medical imaging and segmentation

All 377