Papers

New on arXiv

Tue 13 Jan 2026

348 papers

Do Open-Vocabulary Detectors Transfer to Aerial Imagery? A Comparative EvaluationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Mem2ActBench: A Benchmark for Evaluating Long-Term Memory Utilization in Task-Oriented Autonomous AgentsLLM agents and tool use → Agent memory systems
Agentic Digital Twins: A Taxonomy of Capabilities for Understanding Possible FuturesLLM agents and tool use → General LLM agent frameworks and benchmarks
On the Limits of Learned Importance Scoring for KV Cache CompressionEfficient LLM inference and serving → KV cache optimization
Synthetic Data for Veterinary EHR De-identification: Benefits, Limits, and Safety Trade-offs Under Fixed ComputeSynthetic data and self-improvement for LLMs → Medical synthetic data and agents
The Hierarchy of Agentic Capabilities: Evaluating Frontier Models on Realistic RL EnvironmentsLLM agents and tool use → General LLM agent frameworks and benchmarks
OpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAGRetrieval-augmented generation → RAG robustness and adaptive retrieval
ART: Action-based Reasoning Task Benchmarking for Medical AI AgentsLLM reasoning → Clinical and biomedical applications +1
Breaking the Bottlenecks: Scalable Diffusion Models for 3D Molecular GenerationDiffusion models → Flow matching, time series and science
Imagine-then-Plan: Agent Learning from Adaptive Lookahead with World ModelsWorld models → World models for language agents
LLMs as Assessors: Right for the Right Reason?LLM reasoning → Reasoning benchmarks and evaluation
Attention Consistency Regularization for Interpretable Early-Exit Neural NetworksEfficient LLM inference and serving → Early exit and edge inference
Modeling LLM Agent Reviewer Dynamics in Elo-Ranked Review SystemLLM agents and tool use → General LLM agent frameworks and benchmarks
MemRec: Collaborative Memory-Augmented Agentic Recommender SystemLLM agents and tool use → Agentic recommender systems
Reasoning Matters for 3D Visual GroundingLLM reasoning → Spatial and embodied reasoning
Multiplex Thinking: Reasoning via Token-wise Branch-and-MergeLLM reasoning → Prompting and chain-of-thought reasoning
MixServe: An Automatic Distributed Serving System for MoE Models with Hybrid Parallelism Based on Fused Communication AlgorithmMixture-of-experts → Efficient MoE training and inference systems
Asymptotic Universal Alignment: A New Alignment Framework via Test-Time ScalingLLM reasoning → Reinforcement learning for reasoning +1
Reliable Graph-RAG for Codebases: AST-Derived Graphs vs LLM-Extracted Knowledge GraphsRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Rewarding the Rare: Uniqueness-Aware RL for Creative Problem Solving in LLMsLLM reasoning → Reinforcement learning for reasoning
M3CoTBench: Benchmark Chain-of-Thought of MLLMs in Medical Image UnderstandingLLM reasoning → Clinical and biomedical applications +1
To Retrieve or To Think? Cross-Boundary Context Evolution for Multi-hop Complex ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning +1
TableCache: Primary Foreign Key Guided KV Cache Precomputation for Low Latency Text-to-SQLEfficient LLM inference and serving → KV cache optimization
Inferring Latent Intentions: Attributional Natural Language Inference in LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
From Rows to Reasoning: A Retrieval-Augmented Multimodal Framework for Spreadsheet UnderstandingLLM reasoning → Knowledge- and tool-augmented reasoning +1
PrivGemo: Privacy-Preserving Dual-Tower Graph Retrieval for Empowering LLM Reasoning with Memory AugmentationLLM reasoning → Knowledge- and tool-augmented reasoning
Learning from Demonstrations via Capability-Aware Goal SamplingOffline RL and imitation learning from data → Imitation learning methods
RAGShaper: Eliciting Sophisticated Agentic RAG Skills via Automated Data SynthesisRetrieval-augmented generation → Agentic RAG, agents and memory +2
PersonaDual: Balancing Personalization and Objectivity via Adaptive ReasoningLLM reasoning → Reinforcement learning for reasoning
Advancing ESG Intelligence: An Expert-level Agent and Comprehensive Benchmark for Sustainable FinanceLLM agents and tool use → Financial trading and forecasting agents
Parallel Context-of-Experts Decoding for Retrieval Augmented GenerationRetrieval-augmented generation → RAG robustness and adaptive retrieval +1
VLingNav: Embodied Navigation with Adaptive Reasoning and Visual-Assisted Linguistic MemoryVision-language-action models → Reasoning and chain-of-thought in VLA
Resisting Manipulative Bots in Meme Coin Copy Trading: A Multi-Agent Approach with Chain-of-Thought ReasoningLLM reasoning → Reasoning for agents, planning and code +1
M2FMoE: Multi-Resolution Multi-View Frequency Mixture-of-Experts for Extreme-Adaptive Time Series ForecastingMixture-of-experts → MoE for time series and graphs
SafeRedir: Prompt Embedding Redirection for Robust Unlearning in Image Generation ModelsLLM safety and security → Vision-language and text-to-image safety
GraphSearch: Agentic Search-Augmented Reasoning for Zero-Shot Graph LearningLLM reasoning → Knowledge- and tool-augmented reasoning +1
ViDoRe V3: A Comprehensive Evaluation of Retrieval Augmented Generation in Complex Real-World ScenariosRetrieval-augmented generation → Multimodal RAG
SoC: Semantic Orthogonal Calibration for Test-Time Prompt TuningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
SfMamba: Efficient Source-Free Domain Adaptation via Selective Scan ModelingState-space sequence models (S4, Mamba) → Vision Mamba backbones
ExpSeek: Self-Triggered Experience Seeking for Web AgentsLLM agents and tool use → Agent memory systems
VideoHEDGE: Entropy-Based Hallucination Detection for Video-VLMs via Semantic Clustering and Spatiotemporal PerturbationsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
DeepResearch Bench II: Diagnosing Deep Research Agents via Rubrics from Expert ReportsLLM agents and tool use → Web, search, and research agents
Your Group-Relative Advantage Is BiasedLLM reasoning → Reinforcement learning for reasoning
STAR: Detecting Inference-time Backdoors in LLM Reasoning via State-Transition Amplification RatioLLM reasoning → Reasoning for agents, planning and code +1
STAGE: A Full-Screenplay Benchmark for Reasoning over Evolving StoriesLLM reasoning → Reasoning benchmarks and evaluation
BenchOverflow: Measuring Overflow in Large Language Models via Plain-Text PromptsLLM safety and security → Jailbreaking
Surgical Refusal Ablation: Disentangling Safety from Intelligence via Concept-Guided Spectral CleaningLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Cross-modal Proxy Evolving for OOD Detection with Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Towards Safer Mobile Agents: Scalable Generation and Evaluation of Diverse Scenarios for VLMsMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
JudgeRLVR: Judge First, Generate Second for Efficient ReasoningLLM reasoning → Reinforcement learning for reasoning
Zero-Shot Distracted Driver Detection via Vision Language Models with Double DecouplingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection
CoMa: Contextual Massing Generation with Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
M3-BENCH: Process-Aware Evaluation of LLM Agents' Social Behaviors in Mixed-Motive GamesLLM agents and tool use → Social simulation and multi-agent debate
Decoding Order Matters in Autoregressive Speech SynthesisDiffusion models → Discrete and text diffusion
Beyond Linearization: Attributed Table Graphs for Table ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning
YaPO: Learnable Sparse Activation Steering Vectors for Domain AdaptationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Incentivizing Cardiologist-Like Reasoning in MLLMs for Interpretable Echocardiographic DiagnosisLLM reasoning → Clinical and biomedical applications
Fine-Mem: Fine-Grained Feedback Alignment for Long-Horizon Memory ManagementLLM agents and tool use → Agent memory systems
Large Multimodal Models for Embodied Intelligent Driving: The Next Frontier in Self-Driving?LLM reasoning → Chain-of-thought reasoning for driving +2
Silence the Judge: Reinforcement Learning with Self-Verifier via Latent Geometric ClusteringLLM reasoning → Reinforcement learning for reasoning
Coverage Improvement and Fast Convergence of On-policy Preference LearningLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
MMLGNet: Cross-Modal Alignment of Remote Sensing Data using CLIPMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Hybrid Distillation with CoT Guidance for Edge-Drone Control Code GenerationLLM reasoning → Spatial and embodied reasoning
WebTrap Park: An Automated Platform for Systematic Security Evaluation of Web AgentsLLM agents and tool use → Agent threats, guardrails and governance
Owen-Shapley Policy Optimization: A Principled RL Algorithm for Generative Search LLMsLLM reasoning → Reinforcement learning for reasoning
Controlled LLM Training on Spectral SphereMixture-of-experts → Efficient MoE training and inference systems
Deconstructing Pre-training: Knowledge Attribution Analysis in MoE and Dense ModelsMixture-of-experts → MoE large language models
MMD Guidance: Training-Free Distribution Adaptation for Diffusion Models via Maximum Mean Discrepancy GuidanceDiffusion models → Diffusion foundations and sampling
Scalable Sequential Recommendation under Latency and Memory ConstraintsState-space sequence models (S4, Mamba) → Core SSM architectures for language
Semantic Misalignment in Vision-Language Models under Perceptual DegradationMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
When KV Cache Reuse Fails in Multi-Agent Systems: Cross-Candidate Interaction is Crucial for LLM JudgesLLM agents and tool use → Software engineering coding agents +1
Semantic Laundering in AI Agent Architectures: Why Tool Boundaries Do Not Confer Epistemic WarrantLLM agents and tool use → Agent threats, guardrails and governance
ActiveVLA: Injecting Active Perception into Vision-Language-Action Models for Precise 3D Robotic ManipulationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
AtomMem : Learnable Dynamic Agentic Memory with Atomic Memory OperationLLM agents and tool use → Agent memory systems
Enhancing Image Quality Assessment Ability of LMMs via Retrieval-Augmented GenerationMultimodal LLMs / vision-language models → Image editing and quality assessment +1
ORBIT: On-policy Exploration-Exploitation for Controllable Multi-Budget ReasoningLLM reasoning → Reinforcement learning for reasoning
SnapGen++: Unleashing Diffusion Transformers for Efficient High-Fidelity Image Generation on Edge DevicesDiffusion models → Diffusion foundations and sampling
M3SR: Multi-Scale Multi-Perceptual Mamba for Efficient Spectral ReconstructionState-space sequence models (S4, Mamba) → Hyperspectral image analysis with Mamba
KidVis: Do Multimodal Large Language Models Possess the Visual Perceptual Capabilities of a 6-Year-Old?Multimodal LLMs / vision-language models → MLLM training and visual instruction tuning
D2Plan: Dual-Agent Dynamic Global Planning for Complex Retrieval-Augmented ReasoningLLM reasoning → Knowledge- and tool-augmented reasoning +2
Greedy Is Enough: Sparse Action Discovery in Agentic LLMsLLM agents and tool use → Reinforcement learning for LLM agents
ACE-Router: Generalizing History-Aware Routing from MCP Tools to the Agent WebLLM agents and tool use → General LLM agent frameworks and benchmarks
Discovery and Reinforcement of Tool-Integrated Reasoning Chains via Rollout TreesLLM reasoning → Reasoning for agents, planning and code +1
HIPPO: Accelerating Video Large Language Models Inference via Holistic-aware Parallel Speculative DecodingEfficient LLM inference and serving → Fast autoregressive visual generation
Sparsity Is Necessary: Polynomial-Time Stability for Agentic LLMs in Large Action SpacesLLM agents and tool use → Reinforcement learning for LLM agents
Med-CoReasoner: Reducing Language Disparities in Medical Reasoning via Language-Informed Co-ReasoningLLM reasoning → Clinical and biomedical applications
Diagnosing and Mitigating Sycophancy and Skepticism in LLM Causal JudgmentLLM reasoning → Causal reasoning in LLMs
FSAG: Enhancing Human-to-Dexterous-Hand Finger-Specific Affordance Grounding via Diffusion ModelsDiffusion models → Diffusion and flow policies for robotics
The End of Reward Engineering: How LLMs Are Redefining Multi-Agent CoordinationLLM agents and tool use → Reinforcement learning for LLM agents
MPCI-Bench: A Benchmark for Multimodal Pairwise Contextual Integrity Evaluation of Language Model AgentsLLM agents and tool use → Agent threats, guardrails and governance
User-Oriented Multi-Turn Dialogue Generation with Tool Use at scaleLLM reasoning → Reasoning for agents, planning and code
DNF: Dual-Layer Nested Fingerprinting for Large Language Model Intellectual Property ProtectionLLM safety and security → Watermarking generated content and models
Towards Principled Design of Mixture-of-Experts Language Models under Memory and Inference ConstraintsMixture-of-experts → MoE large language models
Triplets Better Than Pairs: Towards Stable and Effective Self-Play Fine-Tuning for LLMsSynthetic data and self-improvement for LLMs → Self-rewarding alignment and self-training
Evaluating Implicit Regulatory Compliance in LLM Tool Invocation via Logic-Guided SynthesisLLM agents and tool use → Agent threats, guardrails and governance
Route, Retrieve, Reflect, Repair: Self-Improving Agentic Framework for Visual Detection and Linguistic Reasoning in Medical ImagingLLM reasoning → Clinical and biomedical applications +3
ForgetMark: Stealthy Fingerprint Embedding via Targeted Unlearning in Language ModelsLLM safety and security → Watermarking generated content and models
Improving LLM Reasoning with Homophily-aware Structural and Semantic Text-Attributed Graph CompressionLLM reasoning → Knowledge- and tool-augmented reasoning
GI-Bench: A Panoramic Benchmark Revealing the Knowledge-Experience Dissociation of Multimodal Large Language Models in Gastrointestinal Endoscopy Against Clinical StandardsMultimodal LLMs / vision-language models → Medical vision-language models
Prompt-Based Clarity Evaluation and Topic Detection in Political Question AnsweringLLM reasoning → Prompting and chain-of-thought reasoning

All 348