Papers

New on arXiv

Sun 8 Feb 2026

211 papers

LLM Reasoning with Process Rewards for Outcome-Guided StepsLLM reasoning → Reinforcement learning for reasoning +1
Explainable LLM Unlearning Through ReasoningLLM reasoning → Reasoning for agents, planning and code +1
StaTS: Spectral Trajectory Schedule Learning for Adaptive Time Series Forecasting with Frequency Guided DenoiserDiffusion models → Flow matching, time series and science
Can Multimodal LLMs See Science Instruction? Benchmarking Pedagogical Reasoning in K-12 Classroom VideosLLM reasoning → Multimodal reasoning +1
AGORA: Agentic Green Orchestration Architecture for Beyond 5G NetworksLLM agents and tool use → Agentic AI for wireless networks
Evaluating the Impact of Post-Training Quantization on Reliable VQA with Multimodal LLMsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
Can Large Language Models Implement Agent-Based Models? An ODD-based Replication StudyLLM agents and tool use → General LLM agent frameworks and benchmarks
The Confidence Manifold: Geometric Structure of Correctness Representations in Language ModelsMechanistic interpretability → Probing deception, bias and beliefs
Test vs Mutant: Adversarial LLM Agents for Robust Unit Test GenerationLLM agents and tool use → Software engineering coding agents
Robustness of Vision Language Models Against Split-Image Harmful Input AttacksMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +2
Emergent Search and Backtracking in Latent Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
VidVec: Unlocking Video MLLM Embeddings for Video-Text RetrievalMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
Objective Decoupling in Social Reinforcement Learning: Recovering Ground Truth from Sycophantic MajoritiesDeep reinforcement learning methods → RL training of language and agent models
Spectral Guardrails for Agents in the Wild: Detecting Tool Use Hallucinations via Attention TopologyLLM safety and security → Machine unlearning
Reinforcement Learning Method for Zero-Sum Linear-Quadratic Stochastic Differential Games in Infinite HorizonsDeep reinforcement learning methods → RL for games and combinatorial optimization
IRB: Automated Generation of Robust Factuality BenchmarksRetrieval-augmented generation → RAG robustness and adaptive retrieval
ReRoPE: Repurposing RoPE for Relative Camera ControlDiffusion models → Video generation
Enhancing Bandit Algorithms with LLMs for Time-varying User Preferences in Streaming RecommendationsLLM alignment / RLHF / preference optimisation → Preference alignment for recommenders and RAG
Efficient and Adaptable Detection of Malicious LLM Prompts via Bootstrap AggregationSynthetic data and self-improvement for LLMs → Safety of self-evolving systems +1
Compiler-Assisted Speculative Sampling for Accelerated LLM Inference on Heterogeneous Edge DevicesEfficient LLM inference and serving → Speculative decoding
DICE: Disentangling Artist Style from Content via Contrastive Subspace Decomposition in Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Epigraph-Guided Flow Matching for Safe and Performant Offline Reinforcement LearningDiffusion models → Diffusion and flow policies for robotics +1
Graph-Enhanced Deep Reinforcement Learning for Multi-Objective Unrelated Parallel Machine SchedulingDeep reinforcement learning methods → RL for energy, networks and operations
TDGNet: Hallucination Detection in Diffusion Language Models via Temporal Dynamic GraphsDiffusion models → Medical imaging and segmentation
Enhanced Mixture 3D CGAN for Completion and Generation of 3D ObjectsMixture-of-experts → Vision mixture-of-experts
Implicit Strategic Optimization: Rethinking Long-Horizon Decision-Making in Adversarial Poker EnvironmentsLLM reasoning → Reasoning for agents, planning and code
Horizon Imagination: Efficient On-Policy Rollout in Diffusion World ModelsWorld models → Latent dynamics models and model-based RL +1
Free(): Learning to Forget in Malloc-Only Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
Diverge to Induce Prompting: Multi-Rationale Induction for Zero-Shot ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
MIND: Benchmarking Memory Consistency and Action Control in World ModelsWorld models → Video-generation world models and simulators
FlashVID: Efficient Video Large Language Models via Training-free Tree-based Spatiotemporal Token MergingEfficient LLM inference and serving → Visual token pruning in multimodal models
CTFExplorer: Evaluating LLM Offensive Agents Through Multi-Target Web CTF BenchmarkingLLM agents and tool use → Agent threats, guardrails and governance
The Rise of Sparse Mixture-of-Experts: A Survey from Algorithmic Foundations to Decentralized Architectures and Vertical Domain ApplicationsMixture-of-experts → Vision mixture-of-experts
Small Agent Group is the Future of Digital HealthLLM agents and tool use → Medical and scientific domain agents
A Unified Density Operator View of Flow Control and MergingDiffusion models → Diffusion and flow policies for robotics
Towards Adaptive, Scalable, and Robust Coordination of LLM Agents: A Dynamic Ad-Hoc Networking PerspectiveLLM agents and tool use → General LLM agent frameworks and benchmarks
ForecastOcc: Vision-based Semantic Occupancy ForecastingWorld models → World models for autonomous driving
ResidualKV: Residual-Based KV Cache Compression for Efficient Long-Context InferenceEfficient LLM inference and serving → KV cache optimization
Agent Skills: A Data-Driven Analysis of Claude Skills for Extending Large Language Model FunctionalityLLM agents and tool use → Agent skill learning and evolution
Regret Analysis of Unichain Average Reward Constrained MDPs with General ParameterizationDeep reinforcement learning methods → RL theory and sample complexity
Fast Model Selection and Stable Optimization for Softmax-Gated Multinomial-Logistic Mixture of Experts ModelsMixture-of-experts → MoE theory and gating functions
MCIE: Multimodal LLM-Driven Complex Instruction Image Editing with Spatial GuidanceMultimodal LLMs / vision-language models → Image editing and quality assessment
When Is Compositional Reasoning Learnable from Verifiable Rewards?LLM reasoning → Reinforcement learning for reasoning
Analyzing the Impact of Simulation Fidelity on the Evaluation of Autonomous Driving Motion ControlEnd-to-end autonomous driving → Generative simulation for driving
Accelerating Social Science Research via Agentic Hypothesization and ExperimentationLLM agents and tool use → Web, search, and research agents
FSP-Diff: Full-Spectrum Prior-Enhanced DualDomain Latent Diffusion for Ultra-Low-Dose Spectral CT ReconstructionDiffusion models → Medical imaging and segmentation
Cross-Linguistic Persona-Driven Data Synthesis for Robust Multimodal Cognitive Decline DetectionSynthetic data and self-improvement for LLMs → Medical synthetic data and agents
LOCA-bench: Benchmarking Language Agents Under Controllable and Extreme Context GrowthLLM agents and tool use → General LLM agent frameworks and benchmarks
IV Co-Scientist: Multi-Agent LLM Framework for Causal Instrumental Variable DiscoveryLLM agents and tool use → Medical and scientific domain agents
Which private attributes do VLMs agree on and predict well?Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
A Kinetic Energy Perspective of Flow MatchingDiffusion models → Flow matching, time series and science
Selective Fine-Tuning for Targeted and Robust Concept UnlearningDiffusion models → Robustness, privacy and adversarial aspects +1
CausalArmor: Efficient Indirect Prompt Injection Guardrails via Causal AttributionLLM agents and tool use → Agent threats, guardrails and governance +1
MedCoG: Maximizing LLM Inference Density in Medical Reasoning via Meta-Cognitive RegulationLLM reasoning → Clinical and biomedical applications
Rethinking the Value of Agent-Generated Tests for LLM-Based Software Engineering AgentsLLM reasoning → Reasoning for agents, planning and code +1
Rethinking Practical and Efficient Quantization Calibration for Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
Safety Alignment as Continual Learning: Mitigating the Alignment Tax via Orthogonal Gradient ProjectionLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Efficient Anti-exploration via VQVAE and Fuzzy Clustering in Offline Reinforcement LearningOffline RL and imitation learning from data → Offline RL algorithms
MemCoRe: Recovering Evidence from Progressively Compressed Factual Knowledge for Agent MemoryLLM agents and tool use → Agent memory systems
ToolSelf: Unifying Task Execution and Self-Reconfiguration via Tool-Driven Emergent AdaptationLLM agents and tool use → Reinforcement learning for LLM agents
Rethinking Code Complexity Through the Lens of Large Language ModelsLLM reasoning → Reasoning for agents, planning and code
Rethinking Latency Denial-of-Service: Attacking the LLM Serving Framework, Not the ModelEfficient LLM inference and serving → Serving systems and request scheduling
Direct Soft-Policy Sampling via Langevin DynamicsDeep reinforcement learning methods → RL theory and sample complexity
Thinking in Structures: Evaluating Spatial Intelligence in Constraint-Governed SpacesLLM reasoning → Spatial and embodied reasoning
Geometry-Aware Rotary Position Embedding for Consistent Video World ModelWorld models → Video-generation world models and simulators
LQA: A Lightweight Quantized-Adaptive Framework for Vision-Language Models on the EdgeMultimodal LLMs / vision-language models → Visual token pruning and efficiency
MARTI-MARS2: Scaling Multi-Agent Self-Search via Reinforcement Learning for Code GenerationLLM reasoning → Reasoning for agents, planning and code +1
Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative ReasoningLLM reasoning → Spatial and embodied reasoning +1
VFace: A Training-Free Approach for Diffusion-Based Video Face SwappingDiffusion models → Image generation, text-to-image and editing
SPD-Faith Bench: Diagnosing and Improving Faithfulness in Chain-of-Thought for Multimodal Large Language ModelsLLM reasoning → Multimodal reasoning +2
rePIRL: Learn PRM with Inverse RL for LLM ReasoningLLM reasoning → Reinforcement learning for reasoning +1
Time Series Reasoning via Process-Verifiable Thinking Data Synthesis and Scheduling for Tailored LLM ReasoningLLM reasoning → Time series reasoning +1
Out of the box age estimation through facial imagery: A Comprehensive Benchmark of Vision-Language Models vs. out-of-the-box Traditional ArchitecturesMultimodal LLMs / vision-language models → Social bias and fairness in VLMs
VideoTemp-o3: Harmonizing Temporal Grounding and Video Understanding in Agentic Thinking-with-VideosLLM agents and tool use → Multimodal and embodied agents
Fairness Aware Reward OptimizationLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Thinking Is Not Telling: Information Disclosure in User-Service LLM AgentsLLM reasoning → Reasoning for agents, planning and code +1
MaD-Mix: Multi-Modal Data Mixtures via Latent Space Coupling for Vision-Language Model TrainingMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Attn-GS: Attention-Guided Context Compression for Efficient Personalized LLMsEfficient LLM inference and serving → Prompt and context compression
Talk, Judge, Cooperate: Gossip-Driven Indirect Reciprocity in Self-Interested LLM AgentsLLM agents and tool use → Social simulation and multi-agent debate
Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video DiffusionDiffusion models → Video generation
GR2: Generative Reasoning Re-rankerLLM reasoning → LLM reasoning for recommendation
SRR-Judge: Step-Level Rating and Refinement for Enhancing Search-Integrated Reasoning in Search AgentsLLM reasoning → Knowledge- and tool-augmented reasoning
PAND: Prompt-Aware Neighborhood Distillation for Lightweight Fine-Grained Visual ClassificationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Preference Conditioned Multi-Objective Reinforcement Learning: Decomposed, Diversity-Driven Policy OptimizationDeep reinforcement learning methods → RL for games and combinatorial optimization
Learning to Continually Learn via Meta-learning Agentic Memory DesignsLLM agents and tool use → Agent memory systems
Geo-Code: A Code Framework for Reverse Code Generation from Geometric Images Based on Two-Stage Multi-Agent EvolutionLLM reasoning → Spatial and embodied reasoning
HypRAG: Hyperbolic Dense Retrieval for Retrieval Augmented GenerationRetrieval-augmented generation → Generation and context integration
Homophily-aware Supervised Contrastive Counterfactual Augmented Fair Graph Neural Network
UniRank: End-to-End Domain-Specific Reranking of Hybrid Text-Image Candidates
Generative AI in Action: Field Experimental Evidence from Alibaba's Customer Service Operations
Mitigating Forgetting in Continual Learning with Selective Gradient Projection
The Logovista English-Japanese Machine Translation System
Entropic-Time Inference: Self-Organizing Large Language Model Decoding Beyond Attention
Combating data scarcity in recommendation services: Integrating cognitive types of VARK and neural network technologies (LLM)
Old Habits Die Hard: How Conversational History Geometrically Traps LLMs
TopicENA: Enabling Epistemic Network Analysis at Scale through Automated Topic-Based Coding
Token-Oriented Object Notation vs JSON: A Benchmark of Plain and Constrained Decoding Generation
The Hidden Cost of Structured Generation in LLMs: Draft-Conditioned Constrained Decoding
Transforming Science Learning Materials in the Era of Artificial Intelligence
The Landscape of AI in Science Education: What is Changing and How to Respond

All 211