Papers

New on arXiv

Sun 23 Aug 2026

227 papers

RAPID: Reliability-Aware Pair Importance DistillationKnowledge distillation → Knowledge distillation for LLMs
Improving Spatial-Temporal Reasoning in Video-Language Models with Structured Video PromptingLLM reasoning → Multimodal reasoning
Predicting Consequences and Reinforcing Navigation Policies with Latent World ModelsWorld models → World models for navigation
Fusing Perceptual Vision Experts with Multimodal Large Language Models for Explainable Plant Disease Diagnosis: From Benchmark Imagery to Real-World Robotic Field ValidationMultimodal LLMs / vision-language models → Agricultural vision-language models
ToolRobustBench: Stage-Wise Perturbation Evaluation and Failure Diagnosis for Tool-Calling AgentsLLM agents and tool use → Tool learning and function calling
The Blending Ratio Is Not Where the Performance Is: Diagnosing Prototype Blending for Few-Shot Adaptation of Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1
TRACE: Transition-Aware Residual Control for Multi-Objective Materials DiscoveryLLM agents and tool use → Medical and scientific domain agents
Callability Is Not Operability: Controlled Interface Interventions for LLM AgentsLLM agents and tool use → Tool learning and function calling
Physical Agentic AI: An Architecture for Orchestrating a Robot Crew with LLMsLLM agents and tool use → Multimodal and embodied agents
Iteration Without Elaboration: A Simple ReAct Architecture Suffices for Text-to-SQL GenerationRetrieval-augmented generation → RAG robustness and adaptive retrieval
CAI-DLLM: Convergence Aware Inference for Diffusion Language ModelsDiffusion models → Discrete and text diffusion
OmniCAD: A Large-Scale Benchmark for 3D Spatial Reasoning in Robotics AssembliesLLM reasoning → Spatial and embodied reasoning +1
Q-Learning with Stable Infinite-Dimensional Linear Function ApproximationDeep reinforcement learning methods → RL theory and sample complexity
GeoRisk-RAG: A Hierarchy-Aware Risk Framework for Improving RAG Reliability through Selective AnsweringRetrieval-augmented generation → RAG for recommendation
Learning Generalizable Behaviors for Terminal AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Enhancing Sim2Real Transfer for Torque-Controlled Robots through Real2Sim Dynamics Estimation and Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
Teaching LLMs How ICU Physicians Approach Clinical Reasoning Through OMOP-Aligned Retrieval Improves Reasoning Across Clinical DomainsLLM reasoning → Clinical and biomedical applications
DeepSAGE: Stage-Aware Reinforcement Learning for Structured CBT Counseling DialogueDeep reinforcement learning methods → RL training of language and agent models
What AstroPT knows about galaxies, and what that can teach us about LLMsMechanistic interpretability → Sparse autoencoder training and evaluation
NOVA: Technology-Architecture Co-Design of Near-Memory Processing for Attention-SSM-MoE Hybrid LLM InferenceMixture-of-experts → Efficient MoE training and inference systems
Coalition-Aware Skill Reliability for Self-Evolving AgentsLLM agents and tool use → Agent skill learning and evolution +1
Sparse Additive Off-Policy Evaluation for Reinforcement Learning with Potentially Limited Number of TrajectoriesDeep reinforcement learning methods → RL theory and sample complexity
WorldToken: Time-First Sequence Modeling for Robotic Imitation LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
Vision-Language Models for Occupational Physical Exposure Assessment: Estimating External Hand Forces in Manual Material Handling Tasks from RGB VideoMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Weakly supervised concept Bottleneck Learning for Robust Two stage Object centric visual reasoningLLM reasoning → Multimodal reasoning
CausalCache: Conditional High-Fidelity Restoration for Long-Horizon GUI AgentsLLM agents and tool use → GUI and mobile agents
From Diagnosis to Redesign: Using Quantitative Ethnography to Improve Multi-Agent LLM ReasoningLLM reasoning → Reasoning for agents, planning and code +1
BLADE: Bilevel Low-rank Augmented-Lagrangian Erasure for LLM UnlearningLLM safety and security → Machine unlearning
Scaling Curriculum Learning For Autonomous DrivingEnd-to-end autonomous driving → Reinforcement learning for driving +1
STAGE: Stateful Translation to Agentic Graph Execution with Policy-Scoped Context and Deterministic ControlLLM agents and tool use → Agent threats, guardrails and governance
CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model TrajectoriesLLM agents and tool use → Agent memory systems +1
SymmAdapt: Symmetrical Flow Matching for Source-Free Domain Adaptation in Medical Image SegmentationDiffusion models → Medical imaging and segmentation
Stress Testing Unlearning AlgorithmsLLM safety and security → Machine unlearning
RS3-Prune: Read-Sparse, Store-Sparse Token Pruning for Video Object SegmentationEfficient LLM inference and serving → Token pruning and merging for ViTs
VISTA: Test-Time Compositional Alignment for Visual Autoregressive GenerationDiffusion models → Image generation, text-to-image and editing
Who Pays More for Safety? Measuring the Disparate Cost of Safety Alignment across LanguagesLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
HeatTok: Enhancing Remote Sensing Image Understanding via Thermodiffusion-based TokenizationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
GTA-RAG: Graph-Trajectory-Augmented Reinforcement Learning for Multi-Turn Retrieval-Augmented ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning +1
Arbitrage-Aware Multi-Step Forecasting of Implied Volatility Surfaces: Modelling Surface Trajectories Using Latent DiffusionDiffusion models → Flow matching, time series and science
Small Reasoning Models are Instruction Followers in Function CallingLLM reasoning → Test-time compute and efficient thinking +1
Aligned Alone, Misaligned Together: Forecasting Adversarial Capture in LLM Agent PopulationsLLM agents and tool use → Agent threats, guardrails and governance
Think with Structured Grounding: Perceptual Reinforcement Learning for Chart and Visual-Tabular UnderstandingLLM reasoning → Multimodal reasoning +1
Where World Models Break: Natural-Input Failure DiscoveryWorld models → World models for language agents
Robust Bimanual Vision-Language-Action Models via Embarrassingly Simple Modality MaskingVision-language-action models → Human-video and latent-action pretraining
LD4WAM: Learning Latent Dynamics from Human Videos for World Action ModelsWorld models → Embodied and robot world models
LiST: Local-Simplex Test-Time LoRA FusionMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Context-Aware Cluster Decoding: Semantic Anchor-Driven Coherence in dMLLMsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
When Do VLMs Help Arabic Manuscript OCR? A Cross-Dataset StudyMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
WAM-OPD: On-Policy Distillation for World Action ModelsWorld models → Embodied and robot world models +1
Tracing the Unlabeled Storm: Cross-Variable Transfer in a Lagrangian Atmospheric JEPA FrameworkWorld models → Latent predictive world models
When Not to Imitate: Boundary-Aware Skill Memory for Reliable Tool-Use LLM AgentsLLM agents and tool use → Agent skill learning and evolution
Register Shifts Break LLM Safety: A Bengali Benchmark with Culturally Grounded HarmsLLM safety and security → Safety alignment, refusal and guardrails
Mechanistic Interpretability of Chain-of-Thought Reasoning via Sequential Activation PatchingLLM reasoning → Prompting and chain-of-thought reasoning +1
Noise Floor Audit for Agent BenchmarksLLM agents and tool use → Tool learning and function calling
MedReaMM: Evaluating Large Multimodal Models on Expert-Level Clinical Diagnostic SynthesisMultimodal LLMs / vision-language models → Medical vision-language models
Text-Anchored Semantic Perturbations for Transferable Jailbreak Attacks on Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +2
HERO: Human-profile Enhanced Retrieval Optimization Framework for Long-term Agent MemoryLLM agents and tool use → Agent memory systems
StocBench: A Benchmark for Generative Modeling of Stochastic DynamicsDiffusion models → Flow matching, time series and science
Beyond Instance Slots: Semantically Rich World Models for Physical Interaction PlanningWorld models → Embodied and robot world models
Learning from the Test: Self-Referential Differential Testing for Deep RL AgentsOffline RL and imitation learning from data → Safe offline RL +1
OVIBench: Benchmarking Online Video Question Answering under InterruptionMultimodal LLMs / vision-language models → Video understanding with MLLMs
DreamMimic: Learning Visuomotor Whole-Body Loco-Manipulation via World ModelWorld models → Humanoid and legged locomotion world models
Length-Adaptive Decoding for Masked Diffusion Machine TranslationDiffusion models → Discrete and text diffusion
GAN-Diff : Coupling Pretrained WGAN-GP Features with Conditional Diffusion U-NetsDiffusion models → Diffusion foundations and sampling
Clarify User Expertise: Towards Proactive Conversational Agents Tailoring Responses to User ProficiencyLLM agents and tool use → General LLM agent frameworks and benchmarks
Training-Free VLM Personalization via Calibrated Residual DecodingMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Beyond Over-Refusal: Defending Indirect Prompt Injection via Latent Instruction ManifoldsLLM safety and security → Prompt injection and agent security
N"urnberg NLP @ GermEval Shared Task 2026: Harmful Content Detection in German Social Media through Error-Independent LLM VotersLLM safety and security → Safety alignment, refusal and guardrails
Improving Few-Step Language Flows with Untied Self-ConditioningDiffusion models → Discrete and text diffusion
Beyond What Meets the Eye: Unveiling Situational Illusions for Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Prompt-Based Abstention Fails Under Misleading Context: A Controlled Study of Small Frozen RAG ModelsRetrieval-augmented generation → RAG robustness and adaptive retrieval
Risk-Sensitive Reinforcement Learning with Smoothed Quantile ObjectivesDeep reinforcement learning methods → RL theory and sample complexity
PURA: Provably Unbiased and Robust Multi-Bit Watermarking for AI-Generated Text AttributionLLM safety and security → Watermarking generated content and models
UR2-MLLM: Uncertainty-aware Revisit Reasoning in Multimodal Large Language Models for Radiology Report GenerationLLM reasoning → Clinical and biomedical applications +1
Dual-Layer Agentic Memory with Fast Write Routing and Slow ConsolidationLLM agents and tool use → Agent memory systems
On the Capability Separation Between World-Model Policy Learning and Imitated World-Action ModelsWorld models → Embodied and robot world models
How Agents Represent Humans: Human-Directed Stereotypes in an Open Agent Social NetworkLLM agents and tool use → Social simulation and multi-agent debate
Disagree to Explore, Agree to Commit: Routing-Guided Test-Time Scaling for Software AgentsLLM reasoning → Reasoning for agents, planning and code +1
AudioNoisePrints: Model-free audio watermarking using spatial correlation in flow matching TTSDiffusion models → Robustness, privacy and adversarial aspects
Token-Level Likelihood-Array Regression for Membership Inference and AI-Generated Text DetectionLLM safety and security → Membership inference attacks
Role-Specialized Mixture-of-Agents with Open-Weight LLMs for Clinical PredictionLLM agents and tool use → Medical and scientific domain agents
MCP-Universe RL: A Framework for Training MCP Tool-Use Agents via Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents
The Collaboration Tax: How Much LLM Multi-Agent Systems Pay to CoordinateLLM agents and tool use → Social simulation and multi-agent debate
Linear RNN Scaling Laws: When Longer Sequences Beat More Sequences
I-SplineFlow: Learning Monotone Spline Stochastic Interpolant Schedulers for Few-Step Generation
Local SVD-Entropy Maps as a Complementary Structural Representation for Full-Reference and No-Reference Image Quality Assessment
CS-WCP: Robust Conformal Sets for LLM-Judge Traffic Shifts with Uncertain Group Proportions
When Accuracy Gaps Fail to Certify: Auditing Cross-Domain Recalibration of LLM Judges
Enhancing Multiclass Malware Classification in Resource-Constrained Environments
From Sentiment Classification to Actionable and Responsible Feedback: A Scoping Review and Evidence Map of NLP in Student Evaluation of Teaching, 2015-2026
Block-Level Weight-Space Structure Persists Under Post-Training: An Empirical Study Across LLM Families
Correlation-Aware Structured Pruning for Large Language Models
Hierarchical Bayesian optimization of an aircraft-based multi-agent system-of-systems
Helix-FNO: Spectral-Domain Operator Learning Coupled with a High-Fidelity Mechanistic Model for Fast Surrogate Simulation
ST-Topo GAN: A Motor EEG-to-EMG Decoding Model Matched to Wrist Movement Complexity
Architectural and Regularization Components in Deep Learning Medical Image Registration: Systematic Ablation Study
Understanding Temporal Semantic Stability in Open-Vocabulary UAV Perception through Metric 3D Fusion
The Potential of Haptic Foundation Models
XHotpotQA: A Benchmark for Cross-Lingual Knowledge Composition in Multi-Hop Question Answering
Invocation-Level Reliability of Tool-Using Agents

All 227