Papers

New on arXiv

Mon 10 Nov 2025

428 papers

A survey: Information search time optimization based on RAG (Retrieval Augmentation Generation) chatbotRetrieval-augmented generation → RAG pipeline design and evaluation
How Generative AI Empowers Attackers and Defenders Across the Trust & Safety LandscapeLLM safety and security → Prompt injection and agent security
GroupRank: A Groupwise Paradigm for Effective and Efficient Passage Reranking with LLMsLLM reasoning → Knowledge- and tool-augmented reasoning
Environment-Aware Transfer Reinforcement Learning for Sustainable Beam SelectionDeep reinforcement learning methods → RL for energy, networks and operations
Pre-Attention Expert Prediction and Prefetching for Mixture-of-Experts Large Language ModelsMixture-of-experts → MoE large language models
Continual Learning of Domain Knowledge from Human Feedback in Text-to-SQLLLM alignment / RLHF / preference optimisation → Process reward models for reasoning
Mamba-driven multi-perspective structural understanding for molecular ground-state conformation predictionState-space sequence models (S4, Mamba) → Core SSM architectures for language
Predict and Resist: Long-Term Accident Anticipation under Sensor NoiseDeep reinforcement learning methods → RL for autonomous driving and traffic
How Do Data Owners Say No? A Case Study of Data Consent Mechanisms in Web-Scraped Vision-Language AI Training DatasetsLLM safety and security → Watermarking generated content and models
Diffusion Guided Adversarial State Perturbations in Reinforcement LearningDeep reinforcement learning methods → Adversarial robustness and RL in finance
CAPO: Confidence Aware Preference Optimization Learning for Multilingual PreferencesLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
ResearchRubrics: A Benchmark of Prompts and Rubrics For Evaluating Deep Research AgentsLLM agents and tool use → Web, search, and research agents
A Self-Improving Architecture for Dynamic Safety in Large Language ModelsSynthetic data and self-improvement for LLMs → Safety of self-evolving systems
Private-RAG: Answering Multiple Queries with LLMs while Keeping Your Data PrivateRetrieval-augmented generation → Privacy and security of RAG
Partial Action Replacement: Tackling Distribution Shift in Offline MARLOffline RL and imitation learning from data → Offline multi-agent RL
One Router to Route Them All: Homogeneous Expert Routing for Heterogeneous Graph TransformersMixture-of-experts → MoE for time series and graphs
Think Before You Retrieve: Learning Test-Time Adaptive Search with Small Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning
A Decentralized Retrieval Augmented Generation System with Source Reliabilities Secured on BlockchainRetrieval-augmented generation → Privacy and security of RAG
SCALAR: Benchmarking SAE Interaction Sparsity in Toy LLMsMechanistic interpretability → Sparse autoencoder training and evaluation
Procedural Knowledge Improves Agentic LLM WorkflowsLLM agents and tool use → General LLM agent frameworks and benchmarks
LLM Optimization Unlocks Real-Time Pairwise RerankingRetrieval-augmented generation → Generation and context integration
Biologically-Informed Hybrid Membership Inference Attacks on Generative Genomic ModelsLLM safety and security → Membership inference attacks
Toward the Frontiers of Reliable Diffusion Sampling via Adversarial Sinkhorn Attention GuidanceDiffusion models → Diffusion foundations and sampling
Laplacian Score Sharpening for Mitigating Hallucination in Diffusion ModelsDiffusion models → Medical imaging and segmentation
Routing Manifold Alignment Improves Generalization of Mixture-of-Experts LLMsMixture-of-experts → Expert routing strategies
Robot Learning from a Physical World ModelWorld models → Embodied and robot world models
Using Language Models as Closed-Loop High-Level Planners for Robotics Applications: A Brief Overview and BenchmarksMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
SpatialThinker: Reinforcing Scene Graph-Grounded Spatial Reasoning via Dense RewardsLLM reasoning → Spatial and embodied reasoning +1
StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video GenerationDiffusion models → Video generation
Thinking While Speaking: Inference-Time Knowledge Transfer for Responsive and Intelligent Conversational Voice AgentsLLM reasoning → Audio-language model reasoning +1
C3PO: Optimized Large Language Model Cascades with Probabilistic Cost Constraints for ReasoningLLM reasoning → Test-time compute and efficient thinking
Voice-Interactive Surgical Agent for Multimodal Patient Data ControlLLM agents and tool use → Medical and scientific domain agents
A Diffusion Model to Shrink Proteins While Maintaining Their FunctionDiffusion models → Flow matching, time series and science
Teaching Pretrained Language Models to Think Deeper with Retrofitted RecurrenceLLM reasoning → Test-time compute and efficient thinking
Transformers Provably Learn Chain-of-Thought Reasoning with Length GeneralizationLLM reasoning → Prompting and chain-of-thought reasoning
Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-TrainingLLM reasoning → Reinforcement learning for reasoning
Distributional Biases in Post-Training: A Markovian Analysis of Reasoning TrajectoriesLLM reasoning → Reinforcement learning for reasoning
Inference-Time Scaling of Diffusion Models for Infrared Data GenerationDiffusion models → Image generation, text-to-image and editing
DeepPersona: A Generative Engine for Scaling Deep Synthetic PersonasLLM agents and tool use → Social simulation and multi-agent debate
Grounding Computer Use Agents on Human DemonstrationsLLM agents and tool use → GUI and mobile agents
Q-RAG: Long Context Multi-step Retrieval via Value-based Embedder TrainingRetrieval-augmented generation → RAG robustness and adaptive retrieval
IterResearch: Rethinking Long-Horizon Agents with Interaction ScalingLLM agents and tool use → Web, search, and research agents
FinRpt: Dataset, Evaluation System and LLM-based Multi-agent Framework for Equity Research Report GenerationLLM agents and tool use → Financial trading and forecasting agents
JPRO: Automated Multimodal Jailbreaking via Multi-Agent Collaboration FrameworkLLM agents and tool use → Agent threats, guardrails and governance +1
Scalable Decision Making for Games of Imperfect InformationDeep reinforcement learning methods → RL for games and combinatorial optimization
PlanT 2.0: Exposing Biases and Structural Flaws in Closed-Loop DrivingEnd-to-end autonomous driving → End-to-end system architecture
Enabling Off-Policy Imitation Learning with Deep Actor Critic StabilizationOffline RL and imitation learning from data → Imitation learning methods +1
StreamKV: Streaming Video Question-Answering with Segment-based KV Cache Retrieval and CompressionEfficient LLM inference and serving → Visual token pruning in multimodal models
Beyond Detection: Exploring Evidence-based Multi-Agent Debate for Misinformation Intervention and PersuasionLLM agents and tool use → Social simulation and multi-agent debate
PADiff: Predictive and Adaptive Diffusion Policies for Ad Hoc TeamworkDiffusion models → Diffusion and flow policies for robotics
MVU-Eval: Towards Multi-Video Understanding Evaluation for Multimodal LLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger OptimizationLLM safety and security → Backdoor and data poisoning attacks
Graph Representation-based Model Poisoning on the Heterogeneous Internet of AgentsLLM agents and tool use → Agent threats, guardrails and governance
Federated Learning for Video Violence Detection: Complementary Roles of Lightweight CNNs and Vision-Language Models for Energy-Efficient UseMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
AdaRec: Adaptive Recommendation with LLMs via Narrative Profiling and Dual-Channel ReasoningLLM reasoning → LLM reasoning for recommendation
Guiding Generative Models to Uncover Diverse and Novel Crystals via Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
Conditional Diffusion as Latent Constraints for Controllable Symbolic Music GenerationDiffusion models → Audio, speech and music
Dynamics-Decoupled Trajectory Alignment for Sim-to-Real Transfer in Reinforcement Learning for Autonomous DrivingEnd-to-end autonomous driving → Reinforcement learning for driving
Think Consistently, Reason Efficiently: Energy-Based Calibration for Implicit Chain-of-ThoughtLLM reasoning → Prompting and chain-of-thought reasoning
BridgeVoC: Revitalizing Neural Vocoder from a Restoration PerspectiveDiffusion models → Audio, speech and music
More Agents Improve Math Problem Solving but Adversarial Robustness Gap PersistsLLM agents and tool use → Educational agents and math problem solving
Two Heads are Better than One: Distilling Large Language Model Features Into Small Models with Feature Decomposition and MixtureKnowledge distillation → Knowledge distillation for LLMs
MENTOR: A Metacognition-Driven Self-Evolution Framework for Uncovering and Mitigating Implicit Domain Risks in LLMsLLM safety and security → Safety alignment, refusal and guardrails
HENet++: Hybrid Encoding and Multi-task Learning for 3D Perception and End-to-end Autonomous DrivingEnd-to-end autonomous driving → VLMs for autonomous driving
GEWDiff: Geometric Enhanced Wavelet-based Diffusion Model for Hyperspectral Image Super-resolutionDiffusion models → Inverse problems and image restoration
Agentic AI Sustainability Assessment for Supply Chain Document InsightsLLM agents and tool use → General LLM agent frameworks and benchmarks
LLM Driven Processes to Foster Explainable AILLM agents and tool use → General LLM agent frameworks and benchmarks
HDCNet: A Hybrid Depth Completion Network for Grasping Transparent and Reflective ObjectsState-space sequence models (S4, Mamba) → Vision Mamba backbones
Wasm: A Pipeline for Constructing Structured Arabic Interleaved Multimodal CorporaMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
RedOne 2.0: Rethinking Domain-specific LLM Post-Training in Social Networking ServicesLLM reasoning → Reasoning benchmarks and evaluation
ClusterMine: Robust Label-Free Visual Out-Of-Distribution Detection via Concept Mining from Text CorporaMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
RaLD: Generating High-Resolution 3D Radar Point Clouds with Latent DiffusionDiffusion models → 3D, point cloud and scene generation
Improving Deepfake Detection with Reinforcement Learning-Based Adaptive Data AugmentationDeep reinforcement learning methods → RL with diffusion and flow models
From Pretrain to Pain: Adversarial Vulnerability of Video Foundation Models Without Task KnowledgeMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Learning Quantized Continuous Controllers for Integer HardwareDeep reinforcement learning methods → Spiking and neuromorphic RL
Evaluating Large Language Models for Anxiety, Depression, and Stress Detection: Insights into Prompting Strategies and Synthetic DataSynthetic data and self-improvement for LLMs → Medical synthetic data and agents
Diffolio: A Diffusion Model for Multivariate Probabilistic Financial Time-Series Forecasting and Portfolio ConstructionDiffusion models → Flow matching, time series and science
PADM: A Physics-aware Diffusion Model for Attenuation CorrectionDiffusion models → Medical imaging and segmentation
Learning to Focus: Prioritizing Informative Histories with Structured Attention Mechanisms in Partially Observable Reinforcement LearningWorld models → World models for language agents +1
HLPD: Aligning LLMs to Human Language Preference for Machine-Revised Text DetectionLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Fine-Tuning Diffusion-Based Recommender Systems via Reinforcement Learning with Reward Function OptimizationDiffusion models → Diffusion recommenders +2
A Closer Look at Knowledge Distillation in Spiking Neural Network TrainingKnowledge distillation → Distillation for spiking neural networks
RPTS: Tree-Structured Reasoning Process Scoring for Faithful Multimodal EvaluationLLM reasoning → Multimodal reasoning +1
On The Presence of Double-Descent in Deep Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
EduGuardBench: A Holistic Benchmark for Evaluating the Pedagogical Fidelity and Adversarial Safety of LLMs as Simulated TeachersLLM safety and security → Safety alignment, refusal and guardrails
A Two-Stage System for Layout-Controlled Image Generation using Large Language Models and Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Reinforcement Learning Framework For Stochastic Optimal Control Problem Under Model UncertaintyDeep reinforcement learning methods → RL theory and sample complexity
Ambiguity-aware Truncated Flow Matching for Ambiguous Medical Image SegmentationDiffusion models → Medical imaging and segmentation
Differentiated Directional Intervention A Framework for Evading LLM Safety AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Distillation Dynamics: Towards Understanding Feature-Based Distillation in Vision TransformersKnowledge distillation → General knowledge distillation methods
MI-to-Mid Distilled Compression (M2M-DC): An Hybrid-Information-Guided-Block Pruning with Progressive Inner Slicing Approach to Model CompressionKnowledge distillation → General knowledge distillation methods
ConsistTalk: Intensity Controllable Temporally Consistent Talking Head Generation with Diffusion Noise SearchDiffusion models → Audio, speech and music
Integrating Reweighted Least Squares with Plug-and-Play Diffusion Priors for Noisy Image RestorationDiffusion models → Inverse problems and image restoration
Controllable Flow Matching for Online Reinforcement LearningDiffusion models → Diffusion and flow policies for robotics +1
Convergence of Actor-Critic Learning for Mean Field Games and Mean Field Control in Continuous SpacesDeep reinforcement learning methods → RL for games and combinatorial optimization
MathSE: Improving Multimodal Mathematical Reasoning via Self-Evolving Iterative Reflection and Reward-Guided Fine-TuningLLM reasoning → Multimodal reasoning +3
AgentSUMO: An Agentic Framework for Interactive Simulation Scenario Generation in SUMO via Large Language ModelsLLM agents and tool use → Multimodal and embodied agents
Beyond Uniform Deletion: A Data Value-Weighted Framework for Certified Machine UnlearningLLM safety and security → Machine unlearning
Cross-Modal Unlearning via Influential Neuron Path Editing in Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
SAFENLIDB: A Privacy-Preserving Safety Alignment Framework for LLM-based Natural Language Database InterfacesLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1

All 428