Papers

New on arXiv

Fri 31 Oct 2025

317 papers

Foundation Models for Trajectory Planning in Autonomous Driving: A Review of Progress and Open ChallengesEnd-to-end autonomous driving → Reasoning VLA and VLM driving policies
DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action ModelsLLM reasoning → Spatial and embodied reasoning +1
Unsupervised Cycle Detection in Agentic ApplicationsLLM agents and tool use → Software engineering coding agents
Self-HarmLLM: Can Large Language Model Harm Itself?LLM safety and security → Jailbreaking
Simulating Misinformation Vulnerabilities With Agent PersonasLLM agents and tool use → Social simulation and multi-agent debate
EncouRAGe: Evaluating RAG Local, Fast, and ReliableRetrieval-augmented generation → RAG pipeline design and evaluation
SELF-REDRAFT: Eliciting Intrinsic Exploration-Exploitation Balance in Test-Time Scaling for Code GenerationLLM reasoning → Reasoning for agents, planning and code
A Step Toward World Models: A Survey on Robotic ManipulationWorld models → Embodied and robot world models
Thinking Like a Student: AI-Supported Reflective Planning in a Theory-Intensive Computer Science CourseLLM reasoning → Reasoning benchmarks and evaluation
Scalable Processing-Near-Memory for 1M-Token LLM Inference: CXL-Enabled KV-Cache Management Beyond GPU LimitsEfficient LLM inference and serving → KV cache optimization
A Technical Exploration of Causal Inference with Hybrid LLM Synthetic DataSynthetic data and self-improvement for LLMs → LLM synthetic data generation
Language Modeling With Factorization MemoryState-space sequence models (S4, Mamba) → Core SSM architectures for language
LongCat-Flash-Omni Technical ReportMixture-of-experts → Multimodal MoE models
Improving the Robustness of Control of Chaotic Convective Flows with Domain-Informed Reinforcement LearningDeep reinforcement learning methods → RL with diffusion and flow models
Object-Aware 4D Human Motion GenerationDiffusion models → Video generation
Neural Transparency: Mechanistic Interpretability Interfaces for Anticipating Model Behaviors for Personalized AIMechanistic interpretability → Probing deception, bias and beliefs
Consistently Simulating Human Personas with Multi-Turn Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
Iterative Foundation Model Fine-Tuning on Multiple RewardsDeep reinforcement learning methods → RL with diffusion and flow models
Diffusion Models at the Drug Discovery Frontier: A Review on Generating Small Molecules versus Therapeutic PeptidesDiffusion models → Flow matching, time series and science
Diffusion LLMs are Natural Adversaries for any LLMLLM safety and security → Jailbreaking
Position: Vibe Coding Needs Vibe Reasoning: Improving Vibe Coding with Formal VerificationLLM reasoning → Reasoning for agents, planning and code
Understanding Code Agent Behaviour: An Empirical Study of Success and Failure TrajectoriesLLM agents and tool use → Software engineering coding agents
EL-MIA: Quantifying Membership Inference Risks of Sensitive Entities in LLMsLLM safety and security → Membership inference attacks
A Retrospect to Multi-prompt Learning across Vision and LanguageMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Deep reinforcement learning for optimal trading with partial informationDeep reinforcement learning methods → Adversarial robustness and RL in finance
From Evidence to Verdict: An Agent-Based Forensic Framework for AI-Generated Image DetectionMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Can SAEs reveal and mitigate racial biases of LLMs in healthcare?Mechanistic interpretability → Probing deception, bias and beliefs
CompAgent: An Agentic Framework for Visual Compliance VerificationLLM agents and tool use → Multimodal and embodied agents
EgoMI: Learning Active Vision and Whole-Body Manipulation from Egocentric Human DemonstrationsOffline RL and imitation learning from data → Robot manipulation imitation learning
FLoC: Facility Location-Based Efficient Visual Token Compression for Long Video UnderstandingLLM reasoning → Multimodal reasoning +2
End-to-End Dexterous Arm-Hand VLA Policies via Shared Autonomy: VR Teleoperation Augmented by Autonomous Hand VLA Policy for Efficient Data CollectionVision-language-action models → Human-video and latent-action pretraining
Cross-fluctuation phase transitions reveal sampling dynamics in diffusion modelsDiffusion models → Diffusion foundations and sampling
GeneFlow: Translation of Single-cell Gene Expression to Histopathological Images via Rectified FlowDiffusion models → Medical imaging and segmentation
LC-Opt: Benchmarking Reinforcement Learning and Agentic AI for End-to-End Liquid Cooling Optimization in Data CentersLLM agents and tool use → Reinforcement learning for LLM agents
Cognitive Alignment in Personality Reasoning: Leveraging Prototype Theory for MBTI InferenceLLM reasoning → Reasoning benchmarks and evaluation
Phased DMD: Few-step Distribution Matching Distillation via Score Matching within SubintervalsDiffusion models → Diffusion foundations and sampling +1
fabric-lib: RDMA Point-to-Point Communication for LLM SystemsMixture-of-experts → Efficient MoE training and inference systems
SpecAttn: Speculating Sparse AttentionEfficient LLM inference and serving → Speculative decoding
Interaction as Intelligence Part II: Asynchronous Human-Agent Rollout for Long-Horizon Task TrainingLLM agents and tool use → General LLM agent frameworks and benchmarks
Validity Is What You NeedLLM agents and tool use → General LLM agent frameworks and benchmarks
BEAT: Visual Backdoor Attacks on VLM-based Embodied Agents via Contrastive Trigger LearningMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +2
VeriMoA: A Mixture-of-Agents Framework for Spec-to-HDL GenerationLLM agents and tool use → Software engineering coding agents
Dual-Stream Diffusion for World-Model Augmented Vision-Language-Action ModelWorld models → Embodied and robot world models +2
Spatial-SSRL: Enhancing Spatial Understanding via Self-Supervised Reinforcement LearningMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
InnovatorBench: Evaluating Agents' Ability to Conduct Innovative LLM ResearchLLM agents and tool use → Web, search, and research agents
MARAG-R1: Beyond Single Retriever via Reinforcement-Learned Multi-Tool Agentic RetrievalLLM reasoning → Knowledge- and tool-augmented reasoning +2
SIGMA: Search-Augmented On-Demand Knowledge Integration for Agentic Mathematical ReasoningLLM reasoning → Mathematical reasoning +1
Interact-RAG: Reason and Interact with the Corpus, Beyond Black-Box RetrievalLLM reasoning → Knowledge- and tool-augmented reasoning +1
Optimal Convergence Analysis of DDPM for General DistributionsDiffusion models → Diffusion foundations and sampling
Data-Efficient Domain Adaptation for LLM-based MT using Contrastive Preference OptimizationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
EBT-Policy: Energy Unlocks Emergent Physical Reasoning CapabilitiesLLM reasoning → Spatial and embodied reasoning +2
TempoBench: Reasoning Execution Without Causal Attribution Is Just SimulationLLM reasoning → Causal reasoning in LLMs
AstuteRAG-FQA: Task-Aware Retrieval-Augmented Generation Framework for Proprietary Data Challenges in Financial Question AnsweringRetrieval-augmented generation → Legal and financial RAG
Context-Gated Cross-Modal Perception with Visual Mamba for PET-CT Lung Tumor SegmentationState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Thought Branches: Interpreting LLM Reasoning Requires ResamplingLLM reasoning → Prompting and chain-of-thought reasoning
NAUTILUS: A Large Multimodal Model for Underwater Scene UnderstandingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Simplex-to-Euclidean Bijections for Categorical Flow MatchingDiffusion models → Flow matching, time series and science
Diffuse Thinking: Exploring Diffusion Language Models as Efficient Thought Proposers for ReasoningLLM reasoning → Diffusion reasoning and test-time scaling +1
VCORE: Variance-Controlled Optimization-based Reweighting for Chain-of-Thought SupervisionLLM reasoning → Test-time compute and efficient thinking
From Pixels to Paths: A Multi-Agent Framework for Editable Scientific IllustrationLLM agents and tool use → Multimodal and embodied agents
GeoFM: Enhancing Geometric Reasoning of MLLMs via Synthetic Data Generation through Formal LanguageLLM reasoning → Spatial and embodied reasoning +2
DeepCompress: A Dual Reward Strategy for Dynamically Exploring and Compressing Reasoning ChainsLLM reasoning → Test-time compute and efficient thinking
Dynamic Affective Memory Management for Personalized LLM AgentsLLM agents and tool use → Agent memory systems
Agentic LLMs for REST API Test Amplification: A Comparative Study Across Cloud ApplicationsLLM agents and tool use → Software engineering coding agents
Atlas-Alignment: Making Interpretability Transferable Across Language ModelsMechanistic interpretability → Knowledge and representations inside LLMs
Modality Alignment across Trees on Heterogeneous Hyperbolic ManifoldsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
On the Equivalence of Optimal Transport Problem and Action Matching with Optimal Vector FieldsDiffusion models → Flow matching, time series and science
Measuring Chain-of-Thought Monitorability Through Faithfulness and VerbosityLLM reasoning → Prompting and chain-of-thought reasoning
Fine-Tuning Open Video Generators for Cinematic Scene Synthesis: A Small-Data Pipeline with LoRA and Wan2.1 I2VDiffusion models → Video generation
ToolScope: An Agentic Framework for Vision-Guided and Long-Horizon Tool UseLLM agents and tool use → Multimodal and embodied agents
ThoughtProbe: Classifier-Guided LLM Thought Space Exploration via Probing RepresentationsLLM reasoning → Test-time compute and efficient thinking
RzenEmbed: Towards Comprehensive Multimodal RetrievalMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Reasoning Models Sometimes Output Illegible Chains of ThoughtLLM reasoning → Prompting and chain-of-thought reasoning
Understanding the Implicit User Intention via Reasoning with Large Language Model for Image EditingLLM reasoning → Multimodal reasoning +1
When AI Trading Agents Compete: Adverse Selection of Meta-Orders by Reinforcement Learning-Based Market MakingDeep reinforcement learning methods → Adversarial robustness and RL in finance
Reinforcement Learning for Long-Horizon Unordered Tasks: From Boolean to Coupled Reward MachinesDeep reinforcement learning methods → Interpretable and logic-guided RL
Versatile and Efficient Medical Image Super-Resolution Via Frequency-Gated MambaState-space sequence models (S4, Mamba) → Mamba for medical image analysis
Can LLMs Help You at Work? A Sandbox for Evaluating LLM Agents in Enterprise EnvironmentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Rethinking Robust Adversarial Concept Erasure in Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects
FOCUS: Efficient Keyframe Selection for Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Why Do Multilingual Reasoning Gaps Emerge in Reasoning Language Models?LLM reasoning → Reasoning benchmarks and evaluation
Enhancing Trustworthy GUI Grounding via Self-Critiqued Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
T3: Test-Time Model Merging in VLMs for Zero-Shot Medical Imaging AnalysisMultimodal LLMs / vision-language models → Medical vision-language models
RegionRAG: Region-level Retrieval-Augmented Generation for Visual Document UnderstandingRetrieval-augmented generation → Multimodal RAG
ECVL-ROUTER: Scenario-Aware Routing for Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Enhancing Spatio-Temporal Zero-shot Action Recognition with Language-driven Description AttributesMultimodal LLMs / vision-language models → Video understanding with MLLMs
FinPos: A Position-Aware Trading Agent System for Real Financial MarketsLLM agents and tool use → Financial trading and forecasting agents
Approximate Diverse k-nearest Neighbor Search in Vector DatabaseRetrieval-augmented generation → Vector search and ANN indexes
MoRE: 3D Visual Geometry Reconstruction Meets Mixture-of-ExpertsMixture-of-experts → Vision mixture-of-experts
Privacy-Aware Continual Self-Supervised Learning on Multi-Window Chest Computed Tomography for Domain-Shift RobustnessKnowledge distillation → Distillation for medical imaging
GUI-Rise: Structured Reasoning and History Summarization for GUI NavigationLLM reasoning → Multimodal reasoning +1
MemeArena: Automating Context-Aware Unbiased Evaluation of Harmfulness Understanding for Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
H2-Cache: A Novel Hierarchical Dual-Stage Cache for High-Performance Acceleration of Generative Diffusion ModelsDiffusion models → Diffusion foundations and sampling
DANCER: Dance ANimation via Condition Enhancement and Rendering with diffusion modelDiffusion models → Video generation
Generating Accurate and Detailed Captions for High-Resolution ImagesMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Hierarchical Fusion of Local and Global Visual Features with Mixture-of-Experts for Remote Sensing Image Scene ClassificationMixture-of-experts → Vision mixture-of-experts
Measuring the Security of Mobile LLM Agents under Adversarial Prompts from Untrusted Third-Party ChannelsLLM agents and tool use → Agent threats, guardrails and governance +1
E-MMDiT: Revisiting Multimodal Diffusion Transformer Design for Fast Image Synthesis under Limited ResourcesDiffusion models → Diffusion foundations and sampling
AI Agents in Drug DiscoveryLLM agents and tool use → Medical and scientific domain agents
AURA: A Reinforcement Learning Framework for AI-Driven Adaptive Conversational SurveysDeep reinforcement learning methods → RL training of language and agent models

All 317