Papers

New on arXiv

Tue 27 Jan 2026

401 papers

Neural Estimation of Pairwise Mutual Information in Masked Discrete Sequence ModelsDiffusion models → Discrete and text diffusion
SRAG: RAG with Structured Data Improves Vector RetrievalRetrieval-augmented generation → RAG pipeline design and evaluation
VDLM: Variable Diffusion LMs via Robust Latent-to-Text RenderingDiffusion models → Discrete and text diffusion
Understanding LLM Failures: A Multi-Tape Turing Machine Analysis of Systematic Errors in Language Model ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Playing With AI: How Do State-Of-The-Art Large Language Models Perform in the 1977 Text-Based Adventure Game Zork?LLM reasoning → Reasoning for agents, planning and code
An Agentic AI Control Plane for 6G Network Slice Orchestration, Monitoring, and TradingLLM agents and tool use → Agentic AI for wireless networks
A Practical Guide to Agentic AI Transition in OrganizationsLLM agents and tool use → Software engineering coding agents
Formulating Reinforcement Learning for Human-Robot Collaboration through Off-Policy EvaluationOffline RL and imitation learning from data → Offline RL algorithms +1
Visual Affect Analysis: Predicting Emotions of Image Viewers with Vision-Language ModelsMultimodal LLMs / vision-language models → Multimodal emotion understanding
What Hard Tokens Reveal: Exploiting Low-confidence Tokens for Membership Inference Attacks against Large Language ModelsLLM safety and security → Membership inference attacks
LEMON: How Well Do MLLMs Perform Temporal Multimodal Understanding on Instructional Videos?Multimodal LLMs / vision-language models → Video understanding with MLLMs
Membership Inference Attacks Against Fine-tuned Diffusion Language ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
A Reinforcement Learning Based Universal Sequence Design for Polar CodesDeep reinforcement learning methods → RL for energy, networks and operations
In-Context Reinforcement Learning From Suboptimal Historical DataDeep reinforcement learning methods → Deep RL algorithms and representations
Beyond Bug Fixes: An Empirical Investigation of Post-Merge Code Quality Issues in Agent-Generated Pull RequestsLLM agents and tool use → Software engineering coding agents
Structural Anchor Pruning: Training-Free Multi-Vector Compression for Visual Document RetrievalMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Are We All Using Agents the Same Way? An Empirical Study of Core and Peripheral Developers Use of Coding AgentsLLM agents and tool use → Software engineering coding agents
Benchmarking Reward Hack Detection in Code Environments via Contrastive AnalysisLLM alignment / RLHF / preference optimisation → Reward hacking and misspecification
Quantization-Aware Distillation for NVFP4 Inference Accuracy RecoveryMultimodal LLMs / vision-language models → Visual token pruning and efficiency
IMRNNs: An Efficient Method for Interpretable Dense Retrieval via Embedding ModulationRetrieval-augmented generation → Generation and context integration
Sparse CLIP: Co-Optimizing Interpretability and Performance in Contrastive LearningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Distributional value gradients for stochastic environmentsDeep reinforcement learning methods → RL theory and sample complexity
VERGE: Formal Refinement and Guidance Engine for Verifiable LLM ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Insight Agents: An LLM-Based Multi-Agent System for Data InsightsLLM agents and tool use → Web, search, and research agents
CiMRAG: CiM-Aware Domain-Adaptive and Noise-Resilient Retrieval-Augmented Generation for Edge-Based LLMsRetrieval-augmented generation → RAG pipeline design and evaluation
Decomposing multimodal embedding spaces with group-sparse autoencodersMechanistic interpretability → Interpreting vision and multimodal models
Teaching LLMs to Ask: Self-Querying Category-Theoretic Planning for Under-Specified ReasoningLLM reasoning → Test-time compute and efficient thinking
Exploring the holographic entropy cone via reinforcement learningDeep reinforcement learning methods → RL for games and combinatorial optimization
Benchmarking LLAMA Model Security Against OWASP Top 10 For LLM ApplicationsLLM safety and security → Jailbreaking
E2HiL: Entropy-Guided Sample Selection for Efficient Real-World Human-in-the-Loop Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
MeanCache: From Instantaneous to Average Velocity for Accelerating Flow Matching InferenceDiffusion models → Flow matching, time series and science
Self-Distillation Enables Continual LearningOffline RL and imitation learning from data → Decision transformers and sequence modeling +1
Identifying and Transferring Reasoning-Critical Neurons: Improving LLM Inference Reliability via Activation SteeringLLM reasoning → Test-time compute and efficient thinking +1
Visual Generation Unlocks Human-Like Reasoning through Multimodal World ModelsWorld models → World models for language agents +1
When Iterative RAG Beats Ideal Evidence: A Diagnostic Study in Scientific Multi-hop Question AnsweringLLM reasoning → Knowledge- and tool-augmented reasoning +1
Revisiting Incremental Stochastic Majorization-Minimization Algorithms with Applications to Mixture of ExpertsMixture-of-experts → MoE theory and gating functions
Unsupervised Learning of Efficient Exploration: Pre-training Adaptive Policies via Self-Imposed GoalsDeep reinforcement learning methods → Deep RL algorithms and representations
Youtu-VL: Unleashing Visual Potential via Unified Vision-Language SupervisionMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
CASTER: Breaking the Cost-Performance Barrier in Multi-Agent Orchestration via Context-Aware Strategy for Task Efficient RoutingLLM agents and tool use → General LLM agent frameworks and benchmarks
Regularized f-Divergence Kernel TestsLLM safety and security → Machine unlearning
Agentic Design Patterns: A System-Theoretic FrameworkLLM agents and tool use → General LLM agent frameworks and benchmarks
Benchmarking Multimodal Large Language Models for Missing Modality Completion in Product CataloguesMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Veri-Sure: A Contract-Aware Multi-Agent Framework with Temporal Tracing and Formal Verification for Correct RTL Code GenerationLLM agents and tool use → Software engineering coding agents
RvB: Automating AI System Hardening via Iterative Red-Blue GamesLLM safety and security → Prompt injection and agent security
Improving Policy Exploitation in Online Reinforcement Learning with Instant Retrospect ActionDeep reinforcement learning methods → Deep RL algorithms and representations
Scalable Exploration for High-Dimensional Continuous Control via Value-Guided FlowDeep reinforcement learning methods → RL theory and sample complexity
Generalizable Multimodal Large Language Model Editing via Invariant Trajectory LearningMultimodal LLMs / vision-language models → Image editing and quality assessment
DSVM-UNet : Enhancing VM-UNet with Dual Self-distillation for Medical Image SegmentationState-space sequence models (S4, Mamba) → Mamba for medical image analysis +1
Video-KTR: Reinforcing Video Reasoning via Key Token AttributionLLM reasoning → Multimodal reasoning +1
A Benchmark for Audio Reasoning Capabilities of Multimodal Large Language ModelsLLM reasoning → Audio-language model reasoning +1
KeepLoRA: Continual Learning with Residual Gradient AdaptationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
One Token Is Enough: Improving Diffusion Language Models with a Sink TokenDiffusion models → Discrete and text diffusion
Focus on What Really Matters in Low-Altitude Governance: A Management-Centric Multi-Modal Benchmark with Implicitly Coordinated Vision-Language Reasoning FrameworkLLM reasoning → Spatial and embodied reasoning
AC^2-VLA: Action-Context-Aware Adaptive Computation in Vision-Language-Action Models for Efficient Robotic ManipulationVision-language-action models → Efficient VLA inference and adaptation
Tracking Drift: Variation-Aware Entropy Scheduling for Non-Stationary Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
R^3: Replay, Reflection, and Ranking Rewards for LLM Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning
Up to 36x Speedup: Mask-based Parallel Inference Paradigm for Key Information Extraction in MLLMsMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Safe Exploration via Policy PriorsDeep reinforcement learning methods → Safe and constrained RL
Explicit Multi-head Attention for Inter-head Interaction in Large Language ModelsEfficient LLM inference and serving → KV cache optimization
ComAgent: Multi-LLM based Agentic AI Empowered Intelligent Wireless NetworksLLM agents and tool use → Agentic AI for wireless networks
From Atoms to Chains: Divergence-Guided Reasoning Curriculum for Unlabeled LLM Domain AdaptationLLM reasoning → Prompting and chain-of-thought reasoning
Toward Architecture-Aware Evaluation Metrics for LLM AgentsLLM agents and tool use → Software engineering coding agents
ScenePilot-4K: A Large-Scale First-Person Dataset and Benchmark for Vision-Language Models in Autonomous DrivingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
MaDiS: Taming Masked Diffusion Language Models for Sign Language GenerationDiffusion models → Sign language production
Putting Privacy to the Test: Introducing Red Teaming for Research Data AnonymizationLLM safety and security → Prompt injection and agent security
Generalizable Equivariant Diffusion Models for Non-Abelian Lattice Gauge TheoryDiffusion models → Diffusion foundations and sampling
"Do I Trust the AI?" Towards Trustworthy AI-Assisted Diagnosis: Understanding User Perception in LLM-Supported ReasoningLLM reasoning → Clinical and biomedical applications
LURE-RAG: Lightweight Utility-driven Reranking for Efficient RAGRetrieval-augmented generation → Generation and context integration
ALRM: Agentic LLM for Robotic ManipulationLLM agents and tool use → Multimodal and embodied agents
Automated Safety Benchmarking: A Multi-agent Pipeline for LVLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Masked Diffusion Generative RecommendationDiffusion models → Diffusion recommenders
Reversible Simplex Supervision with Post-Action Debt Accounting for Goal-Reaching RLDeep reinforcement learning methods → Safe and constrained RL
Cortex-Grounded Diffusion Models for Brain Image GenerationDiffusion models → Medical imaging and segmentation
LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector AlignmentLLM safety and security → Jailbreaking
Dual-Strategy-Enhanced ConBiMamba for Neural Speaker DiarizationState-space sequence models (S4, Mamba) → Speech and audio SSMs
APC-RL: Exceeding Data-Driven Behavior Priors with Adaptive Policy CompositionDeep reinforcement learning methods → Deep RL algorithms and representations
Dynamic Multi-Expert Projectors with Stabilized Routing for Multilingual Speech RecognitionMixture-of-experts → Multimodal MoE models
From Internal Diagnosis to External Auditing: A VLM-Driven Paradigm for Data-Free Online Backdoor DefenseLLM safety and security → Backdoor and data poisoning attacks
KG-CRAFT: Knowledge Graph-based Contrastive Reasoning with LLMs for Enhancing Automated Fact-checkingLLM reasoning → Knowledge- and tool-augmented reasoning
RPO:Reinforcement Fine-Tuning with Partial Reasoning OptimizationLLM reasoning → Reinforcement learning for reasoning
Divergence-Free Diffusion Models for Incompressible Fluid FlowsDiffusion models → Flow matching, time series and science
From Observations to Events: Event-Aware World Model for Reinforcement LearningWorld models → Latent dynamics models and model-based RL +1
Innovator-VL: A Multimodal Large Language Model for Scientific DiscoveryMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
LightSBB-M: Bridging Schr"odinger and Bass for Generative Diffusion ModelingDiffusion models → Diffusion foundations and sampling
Balancing Sustainability And Performance: The Role Of Small-Scale LLMs In Agentic Artificial Intelligence SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Curiosity Driven Knowledge Retrieval for Mobile AgentsLLM agents and tool use → GUI and mobile agents
Formula-One Prompting: A Composable Equation-First Prefix for Applied MathematicsLLM reasoning → Mathematical reasoning
Continuous-time q-learning for Markov regime switching system under Tsallis entropyDeep reinforcement learning methods → RL theory and sample complexity
MetaGen: Self-Evolving Roles and Topologies for Multi-Agent LLM ReasoningLLM reasoning → Reasoning for agents, planning and code +2
Understanding Dominant Themes in Reviewing Agentic AI-authored CodeLLM agents and tool use → Software engineering coding agents
Smoothing the Score Function to Enhance Generalization in Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Model-Free Output Feedback Stabilization via Policy Gradient MethodsDeep reinforcement learning methods → RL for games and combinatorial optimization
Group Distributionally Robust Optimization-Driven Reinforcement Learning for LLM ReasoningLLM reasoning → Reinforcement learning for reasoning
DART: Diffusion-Inspired Speculative Decoding for Fast LLM InferenceEfficient LLM inference and serving → Speculative decoding
DiaDem: Advancing Dialogue Descriptions in Audiovisual Video Captioning for Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Structure-based RNA Design by Step-wise Optimization of Latent Diffusion ModelDiffusion models → Flow matching, time series and science
LLMs Can Unlearn Refusal with Only 1,000 Benign SamplesLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Towards Pixel-Level VLM Perception via Simple Points PredictionMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
RPO-RAG: Aligning Small LLMs with Relation-aware Preference Optimization for Knowledge Graph Question AnsweringLLM reasoning → Knowledge- and tool-augmented reasoning +2
UniPCB: A Unified Vision-Language Benchmark for Open-Ended PCB Quality InspectionMultimodal LLMs / vision-language models → Chart understanding and anomaly detection

All 401