Papers

New on arXiv

Wed 29 Oct 2025

371 papers

Diverse Preference Learning for Capabilities and AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Future of AI Models: A Computational perspective on Model collapseDiffusion models → Discrete and text diffusion +1
FlowMM: Cross-Modal Information Flow Guided KV Cache Merging for Efficient Multimodal Context InferenceEfficient LLM inference and serving → KV cache optimization
Beyond One-Size-Fits-All: Personalized Harmful Content Detection with In-Context LearningLLM safety and security → Safety alignment, refusal and guardrails
SMAGDi: Socratic Multi Agent Interaction Graph Distillation for Efficient High Accuracy ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Approaching Low-Cost Cardiac Intelligence with Semi-Supervised Knowledge DistillationKnowledge distillation → Distillation for medical imaging
LGCC: Enhancing Flow Matching Based Text-Guided Image Editing with Local Gaussian Coupling and Context ConsistencyMultimodal LLMs / vision-language models → Image editing and quality assessment +1
Retrieval-Augmented Multimodal Depression DetectionRetrieval-augmented generation → Medical and biomedical RAG
Generalizing Test-time Compute-optimal Scaling as an Optimizable GraphLLM reasoning → Reasoning for agents, planning and code
Bridging Vision, Language, and Mathematics: Pictographic Character Reconstruction with B'ezier CurvesMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
ScaleCall -- Agentic Tool Calling at Scale for Fintech: Challenges, Methods, and Deployment InsightsLLM agents and tool use → Tool learning and function calling
Latent Domain Prompt Learning for Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Sharpness-Guided Group Relative Policy Optimization via Probability ShapingLLM reasoning → Reinforcement learning for reasoning
Category-Aware Semantic Caching for Heterogeneous LLM WorkloadsEfficient LLM inference and serving → KV cache optimization
Diffusion-Driven Generation of Minimally Preprocessed Brain MRIDiffusion models → Medical imaging and segmentation
SmoothGuard: Defending Multimodal Large Language Models with Noise Perturbation and Clustering AggregationMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision EncodersMultimodal LLMs / vision-language models → Video understanding with MLLMs
Conformal Prediction Beyond the Horizon: Distribution-Free Inference for Policy EvaluationDeep reinforcement learning methods → RL theory and sample complexity
Rethinking Cross-lingual Alignment: Balancing Transfer and Cultural Erasure in Multilingual LLMsLLM reasoning → Reasoning benchmarks and evaluation
Large Language Model-assisted Autonomous Vehicle Recovery from ImmobilizationEnd-to-end autonomous driving → Reasoning VLA and VLM driving policies
PORTool: Importance-Aware Policy Optimization with Rewarded Tree for Multi-Tool-Integrated ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Dual Mixture-of-Experts Framework for Discrete-Time Survival AnalysisMixture-of-experts → Medical and biomedical MoE
Larger Hausdorff Dimension in Scanning Pattern Facilitates Mamba-Based Methods in Low-Light Image EnhancementState-space sequence models (S4, Mamba) → Mamba for image restoration
From Queries to Insights: Agentic LLM Pipelines for Spatio-Temporal Text-to-SQLLLM agents and tool use → Web, search, and research agents
Supervised Reinforcement Learning: From Expert Trajectories to Step-wise ReasoningLLM reasoning → Reinforcement learning for reasoning
NeuronMLP: Efficient LLM Inference via Singular Value Decomposition Compression and Tiling on AWS TrainiumEfficient LLM inference and serving → Serving systems and request scheduling
Brain-IT: Image Reconstruction from fMRI via Brain-Interaction TransformerDiffusion models → Medical imaging and segmentation
SymCode: A Neurosymbolic Approach to Mathematical Reasoning via Verifiable Code GenerationLLM reasoning → Mathematical reasoning
SplitFlow: Flow Decomposition for Inversion-Free Text-to-Image EditingDiffusion models → Image generation, text-to-image and editing
Estimating cognitive biases with attention-aware inverse planningDeep reinforcement learning methods → RL for autonomous driving and traffic
RECAP: Reproducing Copyrighted Data from LLMs Training with an Agentic PipelineLLM agents and tool use → General LLM agent frameworks and benchmarks
Humains-Junior: A 3.8B Language Model Achieving GPT-4o-Level Factual Accuracy by Directed Exoskeleton ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Generative Image Restoration and Super-Resolution using Physics-Informed Synthetic Data for Scanning Tunneling MicroscopyDiffusion models → Inverse problems and image restoration
FinOps Agent -- A Use-Case for IT Infrastructure and Cost OptimizationLLM agents and tool use → Financial trading and forecasting agents
MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiencyLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
π_textttRL: Online RL Fine-tuning for Flow-based Vision-Language-Action ModelsVision-language-action models → Efficient VLA inference and adaptation
Approximating Human Preferences Using a Multi-Judge Learned SystemLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Synthesizing High-Quality Visual Question Answering from Medical Documents with Generator-Verifier LMMsMultimodal LLMs / vision-language models → Medical vision-language models
AAGATE: A NIST AI RMF-Aligned Governance Platform for Agentic AILLM agents and tool use → Agent threats, guardrails and governance
Through the Judge's Eyes: Inferred Thinking Traces Improve Reliability of LLM RatersLLM reasoning → Reasoning benchmarks and evaluation
Debate2Create: Robot Co-design via Multi-Agent LLM DebateLLM agents and tool use → General LLM agent frameworks and benchmarks
Identity Management for Agentic AI: The new frontier of authorization, authentication, and security for an AI agent worldLLM agents and tool use → Agent threats, guardrails and governance
ScaleDiff: Higher-Resolution Image Synthesis via Efficient and Model-Agnostic DiffusionDiffusion models → Diffusion foundations and sampling
An Agentic Framework for Rapid Deployment of Edge AI Solutions in Industry 5.0LLM agents and tool use → General LLM agent frameworks and benchmarks
Beyond Length: Quantifying Long-Range Information for Long-Context LLM Pretraining DataLLM reasoning → Reasoning benchmarks and evaluation
Mixture-of-Experts Operator Transformer for Large-Scale PDE Pre-TrainingMixture-of-experts → MoE theory and gating functions
Metis-SPECS: Decoupling Multimodal Learning via Self-distilled Preference-based Cold StartMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
LISTEN to Your Preferences: An LLM Framework for Multi-Objective SelectionLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
MemEIC: A Step Toward Continual and Compositional Knowledge EditingMultimodal LLMs / vision-language models → Image editing and quality assessment
FreeArt3D: Training-Free Articulated Object Generation using 3D DiffusionDiffusion models → 3D, point cloud and scene generation
Multimodal Spatial Reasoning in the Large Model Era: A Survey and BenchmarksLLM reasoning → Spatial and embodied reasoning
TheraMind: A Strategic and Adaptive Agent for Longitudinal Psychological CounselingLLM agents and tool use → Medical and scientific domain agents
Scaling Latent Reasoning via Looped Language ModelsLLM reasoning → Test-time compute and efficient thinking
What the "Spotless" Mind Remembers: How Knowledge Entanglement Shapes What Leaks After Unlearning in LLMsLLM safety and security → Machine unlearning
The Tool Decathlon: Benchmarking Language Agents for Diverse, Realistic, and Long-Horizon Task ExecutionLLM agents and tool use → General LLM agent frameworks and benchmarks
BambooKG: A Neurobiologically-inspired Frequency-Weight Knowledge GraphRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Retrieval-Augmented Search for Large-Scale Map Collections with ColPaliRetrieval-augmented generation → Multimodal RAG
Robotic Assistant: Completing Collaborative Tasks with Dexterous Vision-Language-Action ModelsVision-language-action models → Human-video and latent-action pretraining
Process-Level Trajectory Evaluation for Environment Configuration in Software Engineering AgentsLLM agents and tool use → Software engineering coding agents
PairUni: Pairwise Training for Unified Multimodal Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Navigation in a Three-Dimensional Urban Flow using Deep Reinforcement LearningDeep reinforcement learning methods → RL for robot navigation and UAV control
Mechanistic Interpretability of RNNs emulating Hidden Markov ModelsMechanistic interpretability → Circuits and mechanistic analysis
ALDEN: Reinforcement Learning for Active Navigation and Evidence Gathering in Long DocumentsMultimodal LLMs / vision-language models → Document understanding and multimodal RAG +1
EHR-R1: A Reasoning-Enhanced Foundational Language Model for Electronic Health Record AnalysisLLM reasoning → Clinical and biomedical applications
Are Language Models Efficient Reasoners? A Perspective from Logic ProgrammingLLM reasoning → Reasoning benchmarks and evaluation
Evaluating the Role of Verifiers in Test-Time Scaling for Legal Reasoning TasksLLM reasoning → Test-time compute and efficient thinking
FARSIQA: Faithful and Advanced RAG System for Islamic Question AnsweringRetrieval-augmented generation → RAG robustness and adaptive retrieval
Don't Blind Your VLA: Aligning Visual Representations for OOD GeneralizationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Counterfactual-based Agent Influence Ranker for Agentic AI WorkflowsLLM agents and tool use → General LLM agent frameworks and benchmarks
Communication and Verification in LLM Agents towards Collaboration under Information AsymmetryLLM agents and tool use → General LLM agent frameworks and benchmarks
Standardization of Psychiatric Diagnoses -- Role of Fine-tuned LLM Consortium and OpenAI-gpt-oss Reasoning LLM Enabled Decision Support SystemLLM reasoning → Clinical and biomedical applications
Using VLM Reasoning to Constrain Task and Motion PlanningLLM reasoning → Spatial and embodied reasoning
Error Bounds and Optimal Schedules for Masked Diffusions with Factorized ApproximationsDiffusion models → Discrete and text diffusion
Off-policy Reinforcement Learning with Model-based Exploration AugmentationDeep reinforcement learning methods → Deep RL algorithms and representations
Zero Reinforcement Learning Towards General DomainsLLM reasoning → Reinforcement learning for reasoning +2
Retrieval Augmented Generation (RAG) for Fintech: Agentic Design and EvaluationRetrieval-augmented generation → Agentic RAG, agents and memory +1
MTIR-SQL: Multi-turn Tool-Integrated Reasoning Reinforcement Learning for Text-to-SQLLLM reasoning → Knowledge- and tool-augmented reasoning +1
Right for the Right Reasons: Avoiding Reasoning Shortcuts via Prototypical Neurosymbolic AILLM reasoning → Reasoning benchmarks and evaluation
Echo-Conditioned Denoising Diffusion Probabilistic Models for Multi-Target Tracking in RF SensingDiffusion models → Diffusion for compression and wireless
Agentic AI: A Comprehensive Survey of Architectures, Applications, and Future DirectionsLLM agents and tool use → General LLM agent frameworks and benchmarks
What Challenges Do Developers Face in AI Agent Systems? An Empirical Study on Stack Overflow & GitHub IssuesLLM agents and tool use → Software engineering coding agents
Small Talk, Big Impact? LLM-based Conversational Agents to Mitigate Passive Fatigue in Conditional Automated DrivingLLM agents and tool use → Multimodal and embodied agents
Improving Temporal Consistency and Fidelity at Inference-time in Perceptual Video Restoration by Zero-shot Image-based Diffusion ModelsDiffusion models → Inverse problems and image restoration
Seeing, Signing, and Saying: A Vision-Language Model-Assisted Pipeline for Sign Language Data Acquisition and Curation from Social MediaMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Serve Programs, Not PromptsEfficient LLM inference and serving → Serving systems and request scheduling
Sim-to-Real Gentle Manipulation of Deformable and Fragile Objects with Stress-Guided Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
CRMWeaver: Building Powerful Business Agent via Agentic RL and Shared MemoriesLLM agents and tool use → Web, search, and research agents
StreamingCoT: A Dataset for Temporal Dynamics and Multimodal Chain-of-Thought Reasoning in Streaming VideoQALLM reasoning → Multimodal reasoning
GAP: Graph-Based Agent Planning with Parallel Tool Use and Reinforcement LearningLLM agents and tool use → General LLM agent frameworks and benchmarks
Dense and Diverse Goal Coverage in Multi Goal Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
Parrot: A Training Pipeline Enhances Both Program CoT and Natural Language CoT for ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
Teaching Sarcasm: Few-Shot Multimodal Sarcasm Detection via Distillation to a Parameter-Efficient StudentKnowledge distillation → Multimodal distillation
Revisiting scalable sequential recommendation with Multi-Embedding Approach and Mixture-of-ExpertsMixture-of-experts → MoE for recommender systems
DIRC-RAG: Accelerating Edge RAG with Robust High-Density and High-Loading-Bandwidth Digital In-ReRAM ComputationRetrieval-augmented generation → RAG pipeline design and evaluation
LangHOPS: Language Grounded Hierarchical Open-Vocabulary Part SegmentationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
MoEntwine: Unleashing the Potential of Wafer-scale Chips for Large-scale Expert Parallel InferenceMixture-of-experts → Efficient MoE training and inference systems
Balanced conic rectified flowDiffusion models → Flow matching, time series and science
ProMediate: A Socio-cognitive framework for evaluating proactive agents in multi-party negotiationLLM agents and tool use → Social simulation and multi-agent debate
FELA: A Multi-Agent Evolutionary System for Feature Engineering of Industrial Event Log DataLLM agents and tool use → General LLM agent frameworks and benchmarks
RAVR: Reference-Answer-guided Variational Reasoning for Large Language ModelsLLM reasoning → Reinforcement learning for reasoning

All 371