Papers

New on arXiv

Sat 7 Mar 2026

196 papers

MURE: Hierarchical Multi-Resolution Encoding via Vision-Language Models for Visual Document RetrievalMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
AutoTool: Automatic Scaling of Tool-Use Capabilities in RL via Decoupled Entropy ConstraintsLLM reasoning → Reinforcement learning for reasoning
Mind the Discriminability Trap in Source-Free Cross-domain Few-shot LearningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Targeted Bit-Flip Attacks on LLM-Based AgentsLLM agents and tool use → Agent threats, guardrails and governance
Turn: A Language for Agentic ComputationLLM agents and tool use → Software engineering coding agents
Hindsight Credit Assignment for Long-Horizon LLM AgentsLLM agents and tool use → Reinforcement learning for LLM agents
SoK: Agentic Retrieval-Augmented Generation (RAG): Taxonomy, Architectures, Evaluation, and Research DirectionsLLM reasoning → Knowledge- and tool-augmented reasoning +2
Scheduling Parallel Optical Circuit Switches for AI TrainingMixture-of-experts → Efficient MoE training and inference systems
N-Tree Diffusion for Long-Horizon Wildfire Risk ForecastingDiffusion models → Flow matching, time series and science
The Yerkes-Dodson Curve for AI Agents: Emergent Cooperation Under Environmental Pressure in Multi-Agent LLM SimulationsLLM agents and tool use → Social simulation and multi-agent debate
VisualScratchpad: Inference-time Visual Concepts Analysis in Vision Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
FinSheet-Bench: From Simple Lookups to Complex Reasoning, Where LLMs Break on Financial SpreadsheetsLLM reasoning → Reasoning for agents, planning and code
Adversarial Latent-State Training for Robust Policies in Partially Observable DomainsDeep reinforcement learning methods → Adversarial robustness and RL in finance
StructSAM: Structure- and Spectrum-Preserving Token Merging for Segment Anything ModelsEfficient LLM inference and serving → Token pruning and merging for ViTs
Seeing the Reasoning: How LLM Rationales Influence User Trust and Decision-Making in Factual Verification TasksLLM reasoning → Reasoning benchmarks and evaluation
Retrieval-Augmented Multi-scale Framework for County-Level Crop Yield Prediction Across Large RegionsRetrieval-augmented generation → Time series and agricultural RAG
AutoResearch-RL: Perpetual Self-Evaluating Reinforcement Learning Agents for Autonomous Neural Architecture DiscoveryDeep reinforcement learning methods → Deep RL algorithms and representations
MAviS: A Multimodal Conversational Assistant For Avian SpeciesMultimodal LLMs / vision-language models → Agricultural vision-language models
Virtual Try-On for Cultural Clothing: A Benchmarking StudyDiffusion models → 3D, point cloud and scene generation
Adaptive Double-Booking Strategy for Outpatient Scheduling Using Multi-Objective Reinforcement LearningDeep reinforcement learning methods → RL for clinical treatment and surgery
Kinematics-Aware Latent World Models for Data-Efficient Autonomous DrivingWorld models → World models for autonomous driving +1
FabricGen: Microstructure-Aware Woven Fabric GenerationDiffusion models → 3D, point cloud and scene generation
Reinforcement Learning for Vehicle-to-Grid Voltage Regulation: Single-Hub to Multi-Hub Coordination with Battery-Aware ConstraintsDeep reinforcement learning methods → RL for energy, networks and operations
Retrieval-Augmented Generation for Predicting Cellular Responses to Gene PerturbationRetrieval-augmented generation → Medical and biomedical RAG
VINO: Video-driven Invariance for Non-contextual Objects via Structural Prior Guided De-contextualizationKnowledge distillation → Detection and dense prediction distillation
CompassDPO: Dynamics-Controlled Direct Preference Optimization for Robust Safety AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks
Lying to Win: Assessing LLM Deception through Human-AI Games and Parallel-World ProbingLLM agents and tool use → Social simulation and multi-agent debate
Re2: Unlocking LLM Reasoning via Reinforcement Learning with Re-solvingLLM reasoning → Reinforcement learning for reasoning
FastSTAR: Spatiotemporal Token Pruning for Efficient Autoregressive Video SynthesisEfficient LLM inference and serving → Fast autoregressive visual generation
Governance Architecture for Autonomous Agent Systems: Threats, Framework, and Engineering PracticeLLM agents and tool use → Agent threats, guardrails and governance +1
FreeFly-Thinking : Aligning Chain-of-Thought Reasoning with Continuous UAV NavigationLLM reasoning → Spatial and embodied reasoning
Retrieving Minimal and Sufficient Reasoning Subgraphs with Graph Foundation Models for Path-aware GraphRAGRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
ACD-U: Asymmetric co-teaching with machine unlearning for robust learning with noisy labelsLLM safety and security → Machine unlearning
RoTri-Diff: A Spatial Robot-Object Triadic Interaction-Guided Diffusion Model for Bimanual ManipulationDiffusion models → Diffusion and flow policies for robotics +1
Improving reasoning at inference time via uncertainty minimisationLLM reasoning → Test-time compute and efficient thinking
Agentic Planning with Reasoning for Image Styling via Offline RLLLM agents and tool use → Multimodal and embodied agents +1
LiveWorld: Simulating Out-of-Sight Dynamics in Generative Video World ModelsWorld models → Video-generation world models and simulators
PDD: Manifold-Prior Diverse Distillation for Medical Anomaly DetectionKnowledge distillation → Distillation for medical imaging
DexKnot: Generalizable Visuomotor Policy Learning for Dexterous Bag-Knotting ManipulationDiffusion models → Diffusion and flow policies for robotics
The Model Knows Which Tokens Matter: Automatic Token Selection via Noise GatingMultimodal LLMs / vision-language models → Visual token pruning and efficiency
More Than 1v1: Human-AI Alignment in Early Developmental Communities with Multimodal LLMsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Deep Expert Injection for Anchoring Retinal VLMs with Domain-Specific KnowledgeMultimodal LLMs / vision-language models → Medical vision-language models
Learning From Failures: Efficient Reinforcement Learning Control with Episodic MemoryDeep reinforcement learning methods → Deep RL algorithms and representations
Vision Language Models Cannot Reason About Physical TransformationLLM reasoning → Spatial and embodied reasoning +1
Efficient Personalized Reranking with Semi-Autoregressive Generation and Online Knowledge DistillationKnowledge distillation → Distillation for recommender systems
AutoUE: Automated Generation of 3D Games in Unreal Engine via Multi-Agent SystemsLLM agents and tool use → Software engineering coding agents
Facial Expression Generation Aligned with Human Preference for Natural Dyadic InteractionLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Exploring the Reasoning Depth of Small Language Models in Software Architecture: A Multidimensional Evaluation Framework Towards Software Engineering 2.0LLM reasoning → Reasoning for agents, planning and code
mAVE: A Watermark for Joint Audio-Visual Generation ModelsLLM safety and security → Watermarking generated content and models
Countdown-Code: A Testbed for Studying The Emergence and Generalization of Reward Hacking in RLVRLLM alignment / RLHF / preference optimisation → Reward hacking and misspecification
Dreamer-CDP: Improving Reconstruction-free World Models Via Continuous Deterministic Representation PredictionWorld models → Latent dynamics models and model-based RL +1
VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics AwarenessMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Entropy-Aware On-Policy Distillation of Language ModelsKnowledge distillation → On-policy and self-distillation for LLMs
CoTJudger: A Graph-Driven Framework for Automatic Evaluation of Chain-of-Thought Efficiency and Redundancy in LRMsLLM reasoning → Test-time compute and efficient thinking
VirtueBench: Evaluating Trustworthiness under Uncertainty in Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
PolypSteer: Counterfactual Endoscopic Synthesis via Training-Free Activation SteeringDiffusion models → Medical imaging and segmentation +1
SODA: Sensitivity-Oriented Dynamic Acceleration for Diffusion TransformerDiffusion models → Diffusion foundations and sampling
Looking Back and Forth: Cross-Image Attention Calibration and Attentive Preference Learning for Multi-Image Hallucination MitigationMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs +1
Self-Supervised Multi-Modal World Model with 4D Space-Time EmbeddingWorld models → Video-generation world models and simulators
SSP: Safety-guaranteed Surgical Policy via Joint Optimization of Behavioral and Spatial ConstraintsOffline RL and imitation learning from data → Safe offline RL +1
Two Frames Matter: A Temporal Attack for Text-to-Video Model JailbreakingLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Enhancing Web Agents with a Hierarchical Memory TreeLLM agents and tool use → Web, search, and research agents
Hit-RAG: Learning to Reason with Long Contexts via Preference AlignmentLLM reasoning → Test-time compute and efficient thinking +3
RESCHED: Rethinking Flexible Job Shop Scheduling from a Transformer-based Architecture with Simplified StatesDeep reinforcement learning methods → RL for energy, networks and operations
Can Safety Emerge from Weak Supervision? A Systematic Analysis of Small Language ModelsLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Mozart: Modularized and Efficient MoE Training on 3.5D Wafer-Scale Chiplet ArchitecturesMixture-of-experts → Efficient MoE training and inference systems
TrajPred: Trajectory-Conditioned Joint Embedding Prediction for Surgical Instrument-Tissue Interaction Recognition in Vision-Language ModelsMultimodal LLMs / vision-language models → Medical vision-language models
AdaGen: Learning Adaptive Policy for Image SynthesisDiffusion models → Diffusion foundations and sampling
Foundational World Models Accurately Detect Bimanual Manipulator FailuresWorld models → Latent dynamics models and model-based RL
Perception-Aware Multimodal Spatial Reasoning from Monocular ImagesLLM reasoning → Spatial and embodied reasoning +1
Diffusion Controller: Framework, Algorithms and ParameterizationDiffusion models → Diffusion and flow policies for robotics
A Systematic Investigation of Document Chunking Strategies and Embedding SensitivityRetrieval-augmented generation → Multimodal RAG
Topology-Aware Reinforcement Learning over Graphs for Resilient Power Distribution NetworksDeep reinforcement learning methods → RL for energy, networks and operations
Learning Quadruped Walking from Seconds of DemonstrationOffline RL and imitation learning from data → Real-world robot offline RL
Post-Training with Policy Gradients: Optimality and the Base Model BarrierDeep reinforcement learning methods → RL theory and sample complexity
Understanding the Performance Plateau in Text-to-Video Retrieval: A Comprehensive Empirical and Linguistic Analysis
RedParrot: Accelerating NL-to-DSL for Business Analytics via Query Semantic Caching
Duration-Informed Workload Scheduler
Large Language Models Unpack Complex Political Opinions through Target-Stance Extraction
Did You Forget What I Asked? Prospective Memory Failures in Large Language Models
Konkani LLM: Multi-Script Instruction Tuning and Evaluation for a Low-Resource Indian Language
Fusing Driver Perceived and Physical Risk for Safety Critical Scenario Screening in Autonomous Driving
Beyond Reward Suppression: Reshaping Steganographic Communication Protocols in MARL via Dynamic Representational Circuit Breaking
Discovering the Hidden Role of Gini Index In Prompt-based Classification
Recursive Language Models Meet Uncertainty: The Surprising Effectiveness of Self-Reflective Program Search for Long Context
Prompt Complexity Dilutes Structured Reasoning: A Follow-Up Study on the Car Wash Problem
Thermal Robustness of Retrieval in Dense Associative Memories: LSE vs LSR Kernels
PolyGLU: State-Conditional Activation Routing in Transformer Feed-Forward Networks
Post Training Quantization for Efficient Dataset Condensation
DDS-UDA: Dual-Domain Synergy for Unsupervised Domain Adaptation in Joint Segmentation of Optic Disc and Optic Cup
DyACE: Dynamic Algorithm Co-evolution for Online Automated Heuristic Design with Large Language Model
AI-Driven Predictive Maintenance with Environmental Context Integration for Connected Vehicles: Simulation, Benchmarking, and Field Validation
MS2MetGAN: Latent-space adversarial training for metabolite-spectrum matching in MS/MS database search
Complementarity-Supervised Spectral-Band Routing for Multimodal Emotion Recognition
The DIME Architecture: A Unified Operational Algorithm for Neural Representation, Dynamics, Control and Integration
AMB-DSGDN: Adaptive Modality-Balanced Dynamic Semantic Graph Differential Network for Multimodal Emotion Recognition
Permutation-Equivariant 2D State Space Models: Theory and Canonical Architecture for Multivariate Time Series
Domain-Specific Quality Estimation for Machine Translation in Low-Resource Scenarios
ConfHit: Conformal Generative Design with Oracle Free Guarantees
Learning to Reflect: Hierarchical Multi-Agent Reinforcement Learning for CSI-Free mmWave Beam-Focusing

All 196