Papers

New on arXiv

Fri 6 Mar 2026

433 papers

StratRAG: A Multi-Hop Retrieval Evaluation Dataset for Retrieval-Augmented Generation SystemsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Your Reviews Replicate You: LLM-Based Agents as Customer Digital Twins for Conjoint AnalysisLLM agents and tool use → Social simulation and multi-agent debate
UniMamba: A Unified Spatial-Temporal Modeling Framework with State-Space and Attention IntegrationState-space sequence models (S4, Mamba) → Time series forecasting with SSMs
Eco-Bee: A Personalised Multi-Modal Agent for Advancing Student Climate Awareness and Sustainable Behaviour in Campus EcosystemsLLM agents and tool use → Social simulation and multi-agent debate
FVRuleLearner: Operator-Level Reasoning Tree (Op-Tree)-Based Rules Learning for Formal VerificationLLM reasoning → Reasoning benchmarks and evaluation
X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMsKnowledge distillation → On-policy and self-distillation for LLMs
The Compression Paradox in LLM Inference: Provider-Dependent Energy Effects of Prompt CompressionEfficient LLM inference and serving → Prompt and context compression
Compression Method Matters: Benchmark-Dependent Output Dynamics in LLM Prompt CompressionEfficient LLM inference and serving → Efficient reasoning and agent inference
Prompt Compression in Production Task Orchestration: A Pre-Registered Randomized TrialEfficient LLM inference and serving → Efficient reasoning and agent inference
Navigating the Concept Space of Language ModelsMechanistic interpretability → SAE features and their applications
Cluster-R1: Large Reasoning Models Are Instruction-following Clustering AgentsLLM reasoning → Reasoning benchmarks and evaluation
Beyond Scalar Rewards: Distributional Reinforcement Learning with Preordered Objectives for Safe and Reliable Autonomous DrivingEnd-to-end autonomous driving → Reinforcement learning for driving +1
Information-Theoretic Constraints for Continual Vision-Language-Action AlignmentVision-language-action models → Reinforcement learning and post-training
Bayesian Conservative Policy Optimization (BCPO): A Novel Uncertainty-Calibrated Offline Reinforcement Learning with Credible Lower BoundsOffline RL and imitation learning from data → Offline RL algorithms
Evaluating Generalization Mechanisms in Autonomous Cyber Attack AgentsLLM agents and tool use → Agent threats, guardrails and governance
Joint MDPs and Reinforcement Learning in Coupled-Dynamics EnvironmentsDeep reinforcement learning methods → RL theory and sample complexity
Swimba: Switch Mamba Model Scales State Space ModelsMixture-of-experts → Efficient MoE training and inference systems +1
Reforming the Mechanism: Editing Reasoning Patterns in LLMs with Circuit ReshapingLLM reasoning → Reasoning benchmarks and evaluation
DLRMamba: Distilling Low-Rank Mamba for Edge Multispectral Fusion Object DetectionState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
A Dynamic Self-Evolving Extraction SystemSynthetic data and self-improvement for LLMs → Self-evolving search agents
XGenBoost: Synthesizing Small and Large Tabular Datasets with XGBoostDiffusion models → Flow matching, time series and science
Empowering Locally Deployable Medical Agent via State Enhanced Logical Skills for FHIR-based Clinical TasksLLM agents and tool use → Medical and scientific domain agents
Distributed Legal Infrastructure for a Trustworthy Agentic WebLLM agents and tool use → General LLM agent frameworks and benchmarks
Not Too Short, Not Too Long: How LLM Response Length Shapes People's Critical Thinking in Error DetectionLLM reasoning → Reasoning benchmarks and evaluation
LieCraft: A Multi-Agent Framework for Evaluating Deceptive Capabilities in Language ModelsLLM agents and tool use → Social simulation and multi-agent debate
PICS: Pairwise Image Compositing with Spatial InteractionsDiffusion models → 3D, point cloud and scene generation
LEAD: Breaking the No-Recovery Bottleneck in Long-Horizon ReasoningLLM reasoning → Test-time compute and efficient thinking
Exact Is Easier: Credit Assignment for Cooperative LLM AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Evaluating Multi-Agent LLM Architectures for Rare Disease DiagnosisLLM agents and tool use → Medical and scientific domain agents
What Does AI Do for Cultural Interpretation? A Randomized Experiment on Close Reading Poems with Exposure to AI InterpretationLLM reasoning → Reasoning benchmarks and evaluation
Are Audio-Language Models Listening? Audio-Specialist Heads for Adaptive Audio SteeringMultimodal LLMs / vision-language models → Video understanding with MLLMs
Characterizing Faults in Agentic AI: A Taxonomy of Types, Symptoms, and Root CausesLLM reasoning → Reasoning for agents, planning and code +1
Step-Level Visual Grounding Faithfulness Predicts Out-of-Distribution Generalization in Long-Horizon Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Breaking the Martingale Curse: Multi-Agent Debate via Asymmetric Cognitive Potential EnergyLLM reasoning → Reasoning for agents, planning and code +1
Best-of-Tails: Bridging Optimism and Pessimism in Inference-Time AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Optimistic Policy RegularizationDeep reinforcement learning methods → RL theory and sample complexity
Physics-Informed Diffusion Model for Generating Synthetic Extreme Rare Weather Events DataDiffusion models → Flow matching, time series and science
HGT-Scheduler: Deep Reinforcement Learning for the Job Shop Scheduling Problem via Heterogeneous Graph TransformersDeep reinforcement learning methods → RL for energy, networks and operations
HybridMimic: Hybrid RL-Centroidal Control for Humanoid Motion MimickingDeep reinforcement learning methods → RL for robot locomotion and manipulation
ButterflyViT: 354× Expert Compression for Edge Vision TransformersMixture-of-experts → Vision mixture-of-experts
Enhancing Instruction Following of LLMs via Activation Steering with Dynamic RejectionMechanistic interpretability → Activation steering methods
Stabilizing Reinforcement Learning for Diffusion Language ModelsLLM reasoning → Reinforcement learning for reasoning +1
Heterogeneous Decentralized Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Agent Hunt: Bounty Based Collaborative Autoformalization With LLM AgentsLLM agents and tool use → Educational agents and math problem solving
Safe Transformer: An Explicit Safety Bit For Interpretable And Controllable AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Multimodal Large Language Models as Image ClassifiersMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Omni-Diffusion: Unified Multimodal Understanding and Generation with Masked Discrete DiffusionMultimodal LLMs / vision-language models → Image editing and quality assessment +1
SUREON: A Benchmark and Vision-Language-Model for Surgical ReasoningLLM reasoning → Clinical and biomedical applications
Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision EncodersMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Boosting deep Reinforcement Learning using pretraining with Logical OptionsDeep reinforcement learning methods → Interpretable and logic-guided RL
NEGATE: Constrained Semantic Guidance for Linguistic Negation in Text-to-Video DiffusionDiffusion models → Discrete and text diffusion
When One Modality Rules Them All: Backdoor Modality Collapse in Multimodal Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
Self-Supervised Flow Matching for Scalable Multi-Modal SynthesisDiffusion models → Flow matching, time series and science
Beyond Rows to Reasoning: Agentic Retrieval for Multimodal Spreadsheet Understanding and EditingLLM reasoning → Knowledge- and tool-augmented reasoning +2
COLD-Steer: Steering Large Language Models via In-Context One-step Learning DynamicsMechanistic interpretability → Activation steering methods
GreenRFM: Learning a resource-efficient radiology vision-language foundation model via supervision-centric pre-trainingMultimodal LLMs / vision-language models → Medical vision-language models
Do Foundation Models Know Geometry? Probing Frozen Features for Continuous Physical MeasurementMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Training Flow Matching: The Role of Weighting and ParameterizationDiffusion models → Flow matching, time series and science
What if? Emulative Simulation with World Models for Situated ReasoningWorld models → World models for language agents +1
Abductive Reasoning with Syllogistic Forms in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
From Prompting to Preference Optimization: A Comparative Study of LLM-based Automated Essay ScoringLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Evaluation of Deontic Conditional Reasoning in Large Language Models: The Case of Wason's Selection TaskLLM reasoning → Reasoning benchmarks and evaluation
A Reference Architecture of Reinforcement Learning FrameworksDeep reinforcement learning methods → Deep RL algorithms and representations
Physical Simulator In-the-Loop Video GenerationDiffusion models → Video generation
Adapter-Augmented Bandits for Online Multi-Constrained Multi-Modal Inference SchedulingMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
DiffInf: Influence-Guided Diffusion for Supervision Alignment in Facial Attribute LearningDiffusion models → Image generation, text-to-image and editing
Talk Freely, Execute Strictly: Schema-Gated Agentic AI for Flexible and Reproducible Scientific WorkflowsLLM agents and tool use → Software engineering coding agents
MoEMambaMIL: Structure-Aware Selective State Space Modeling for Whole-Slide Image AnalysisMixture-of-experts → Medical and biomedical MoE +1
OralGPT-Plus: Learning to Use Visual Tools via Reinforcement Learning for Panoramic X-ray AnalysisLLM reasoning → Clinical and biomedical applications +1
ESAA-Security: An Event-Sourced, Verifiable Architecture for Agent-Assisted Security Audits of AI-Generated CodeLLM agents and tool use → Agent threats, guardrails and governance
LATO: 3D Mesh Flow Matching with Structured TOpology Preserving LAtentsDiffusion models → 3D, point cloud and scene generation
DC-DiT: Adaptive Compute and Elastic Inference for Visual Generation via Dynamic ChunkingDiffusion models → Diffusion foundations and sampling
MoEless: Efficient MoE LLM Serving via Serverless ComputingMixture-of-experts → Efficient MoE training and inference systems +1
K-MaT: Knowledge-Anchored Manifold Transport for Cross-Modal Prompt Learning in Medical ImagingMultimodal LLMs / vision-language models → Medical vision-language models
WorldCache: Accelerating World Models for Free via Heterogeneous Token CachingWorld models → Latent predictive world models
From Entropy to Calibrated Uncertainty: Training Language Models to Reason About UncertaintyLLM reasoning → Test-time compute and efficient thinking
WMoE-CLIP: Wavelet-Enhanced Mixture-of-Experts Prompt Learning for Zero-Shot Anomaly DetectionMultimodal LLMs / vision-language models → Chart understanding and anomaly detection +1
DEX-AR: A Dynamic Explainability Method for Autoregressive Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
3D CBCT Artefact Removal Using Perpendicular Score-Based Diffusion ModelsDiffusion models → Medical imaging and segmentation
The EpisTwin: A Knowledge Graph-Grounded Neuro-Symbolic Architecture for Personal AIRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Artificial Intelligence for Climate Adaptation: Reinforcement Learning for Climate Change-Resilient TransportDeep reinforcement learning methods → RL for energy, networks and operations
Spectral and Trajectory Regularization for Diffusion Transformer Super-ResolutionDiffusion models → Inverse problems and image restoration
Stem: Rethinking Causal Information Flow in Sparse AttentionLLM reasoning → Test-time compute and efficient thinking
Agentic retrieval-augmented reasoning reshapes collective reliability under model variability in radiology question answeringLLM reasoning → Clinical and biomedical applications +2
HiPP-Prune: Hierarchical Preference-Conditioned Structured Pruning for Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Mind the Gap: Pitfalls of LLM Alignment with Asian Public OpinionLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
GazeMoE: Perception of Gaze Target with Mixture-of-ExpertsMixture-of-experts → Vision mixture-of-experts
NOVA: Next-step Open-Vocabulary Autoregression for 3D Multi-Object Tracking in Autonomous DrivingEnd-to-end autonomous driving → VLMs for autonomous driving
Synthetic Monitoring Environments for Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
MLLMRec-R1: Incentivizing Reasoning Capability in Large Language Models for Multimodal Sequential RecommendationLLM reasoning → LLM reasoning for recommendation
SPOT: Span-level Pause-of-Thought for Efficient and Interpretable Latent Reasoning in Large Language ModelsLLM reasoning → Test-time compute and efficient thinking
Conversational Demand Response: Bidirectional Aggregator-Prosumer Coordination through Agentic AILLM agents and tool use → Financial trading and forecasting agents
LIT-RAGBench: Benchmarking Generator Capabilities of Large Language Models in Retrieval-Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
Contrastive-to-Self-Supervised: A Two-Stage Framework for Script Similarity LearningKnowledge distillation → Knowledge distillation for LLMs
Making Training-Free Diffusion Segmentors Scale with the Generative PowerDiffusion models → Medical imaging and segmentation
Optimizing 3D Diffusion Models for Medical Imaging via Multi-Scale Reward LearningDiffusion models → Medical imaging and segmentation +1
Reflective Flow Sampling EnhancementDiffusion models → Image generation, text-to-image and editing
Dual-Agent Multiple-Model Reinforcement Learning for Event-Triggered Human-Robot Co-Adaptation in Decoupled Task SpacesDeep reinforcement learning methods → RL for robot locomotion and manipulation
VLM-RobustBench: A Comprehensive Benchmark for Robustness of Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Spatial Colour Mixing Illusions as a Perception Stress Test for Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

All 433