Papers

New on arXiv

Thu 26 Jun 2025

331 papers

The Value of Gen-AI Conversations: A bottom-up Framework for AI Value AlignmentLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Theory of Mind in Action: The Instruction Inference Task in Dynamic Human-Agent CollaborationLLM agents and tool use → General LLM agent frameworks and benchmarks
Large Language Model Agent for Modular Task Execution in Drug DiscoveryLLM agents and tool use → Medical and scientific domain agents
Leveraging Unlabeled Audio-Visual Data in Speech Emotion Recognition using Knowledge DistillationKnowledge distillation → Multimodal distillation
Red Teaming for Generative AI, Report on a Copyright-Focused Exercise Completed in an Academic Medical CenterLLM safety and security → Prompt injection and agent security
Weak-to-Strong GraphRAG: Aligning Weak Retrievers with Large Language Models for Graph-based Retrieval Augmented GenerationRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
FreeDNA: Endowing Domain Adaptation of Diffusion-Based Dense Prediction with Training-Free Domain Noise AlignmentDiffusion models → Diffusion foundations and sampling
CitySim: Modeling Urban Behaviors and City Dynamics with Large-Scale LLM-Driven Agent SimulationLLM agents and tool use → Multimodal and embodied agents
Why Neural Network Can Discover Symbolic Structures with Gradient-based Training: An Algebraic and Geometric Foundation for Neurosymbolic ReasoningLLM reasoning → Reasoning benchmarks and evaluation
M3PO: Massively Multi-Task Model-Based Policy OptimizationDeep reinforcement learning methods → RL theory and sample complexity
Gradient-Based Neuroplastic Adaptation for Concurrent Optimization of Neuro-Fuzzy NetworksDeep reinforcement learning methods → Deep RL algorithms and representations
THE-Tree: Can Tracing Historical Evolution Enhance Scientific Verification and Reasoning?LLM reasoning → Knowledge- and tool-augmented reasoning
ViStruct: Simulating Expert-Like Reasoning Through Task Decomposition and Visual Attention CuesLLM reasoning → Multimodal reasoning
TADA: Improved Diffusion Sampling with Training-free Augmented DynamicsDiffusion models → Diffusion foundations and sampling
Inverse Design of Diffractive Metasurfaces Using Diffusion ModelsDiffusion models → Flow matching, time series and science
Critically-Damped Higher-Order Langevin Dynamics for Generative ModelingDiffusion models → Diffusion foundations and sampling
Hierarchical Reasoning ModelLLM reasoning → Test-time compute and efficient thinking
Experimental investigation of pose informed reinforcement learning for skid-steered visual navigationDeep reinforcement learning methods → RL for autonomous driving and traffic
Asymmetric Dual Self-Distillation for 3D Self-Supervised Representation LearningKnowledge distillation → Detection and dense prediction distillation
Elucidating and Endowing the Diffusion Training Paradigm for General Image RestorationDiffusion models → Inverse problems and image restoration
CaloHadronic: a diffusion model for the generation of hadronic showersDiffusion models → Flow matching, time series and science
ODEt(ODEl): Shortcutting the Time and the Length in Diffusion and Flow Models for Faster SamplingDiffusion models → Diffusion foundations and sampling
FOCUS: Internal MLLM Representations for Efficient Fine-Grained Visual Question AnsweringLLM reasoning → Multimodal reasoning +1
Risk-Averse Total-Reward Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
TanDiT: Tangent-Plane Diffusion Transformer for High-Quality 360deg Panorama GenerationDiffusion models → 3D, point cloud and scene generation
SEEA-R1: Tree-Structured Reinforcement Fine-Tuning for Self-Evolving Embodied AgentsLLM agents and tool use → Multimodal and embodied agents +1
Fine-Grained Preference Optimization Improves Spatial Reasoning in VLMsLLM reasoning → Spatial and embodied reasoning +2
APO: Enhancing Reasoning Ability of MLLMs via Asymmetric Policy OptimizationLLM reasoning → Reinforcement learning for reasoning +1
Grokking in LLM Pretraining? Monitor Memorization-to-Generalization without TestMixture-of-experts → MoE large language models
Counterfactual Segmentation Reasoning: Diagnosing and Mitigating Pixel-Grounding HallucinationLLM reasoning → Hallucination detection and mitigation +1
StruMamba3D: Exploring Structural Mamba for Self-supervised Point Cloud Representation LearningState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
WorldVLA: Towards Autoregressive Action World ModelWorld models → Embodied and robot world models
Exploring the Design Space of 3D MLLMs for CT Report GenerationMultimodal LLMs / vision-language models → Medical vision-language models
Curing Semantic Drift: A Dynamic Approach to Grounding Generation in Large Vision-Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Mind2Web 2: Evaluating Agentic Search with Agent-as-a-JudgeLLM agents and tool use → Web, search, and research agents
Enhancing User Engagement in Socially-Driven Dialogue through Interactive LLM AlignmentsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
SmoothSinger: A Conditional Diffusion Model for Singing Voice Synthesis with Multi-Resolution ArchitectureDiffusion models → Audio, speech and music
Global and Local Entailment Learning for Natural World ImageryMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
TopK Language ModelsMechanistic interpretability → SAE features and their applications
Optimising 4th-Order Runge-Kutta Methods: A Dynamic Heuristic Approach for Efficiency and Low StorageDeep reinforcement learning methods → RL with diffusion and flow models
Aligning Spoken Dialogue Models from User InteractionsLLM alignment / RLHF / preference optimisation → Preference alignment for recommenders and RAG
MindCube: Spatial Mental Modeling from Limited ViewsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Rethinking Oversaturation in Classifier-Free Guidance via Low FrequencyDiffusion models → Diffusion foundations and sampling
ThinkSound: Chain-of-Thought Reasoning in Multimodal Large Language Models for Audio Generation and EditingLLM reasoning → Audio-language model reasoning +1
Controllable 3D Placement of Objects with Scene-Aware Diffusion ModelsDiffusion models → 3D, point cloud and scene generation
Flow-Based Single-Step Completion for Efficient and Expressive Policy LearningDiffusion models → Diffusion and flow policies for robotics +2
XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT ModulationDiffusion models → Image generation, text-to-image and editing
TableMoE: Neuro-Symbolic Routing for Structured Expert Reasoning in Multimodal Table UnderstandingLLM reasoning → Knowledge- and tool-augmented reasoning +1
Leveraging LLM-Assisted Query Understanding for Live Retrieval-Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
rQdia: Regularizing Q-Value Distributions With Image AugmentationDeep reinforcement learning methods → Deep RL algorithms and representations
ShotBench: Expert-Level Cinematic Understanding in Vision-Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Latent Prototype Routing: Achieving Near-Perfect Load Balancing in Mixture-of-ExpertsMixture-of-experts → Expert routing strategies
SimVecVis: A Dataset for Enhancing MLLMs in Visualization UnderstandingMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
LLaVA-Pose: Enhancing Human Pose and Action Understanding via Keypoint-Integrated Instruction TuningMultimodal LLMs / vision-language models → Video understanding with MLLMs
Continual Self-Supervised Learning with Masked Autoencoders in Remote SensingKnowledge distillation → Detection and dense prediction distillation
HieraSurg: Hierarchy-Aware Diffusion Model for Surgical Video GenerationDiffusion models → Medical imaging and segmentation
Double-Checker: Enhancing Reasoning of Slow-Thinking LLMs via Self-Critical Fine-TuningLLM reasoning → Test-time compute and efficient thinking
HumanOmniV2: From Understanding to Omni-Modal Reasoning with ContextLLM reasoning → Multimodal reasoning +1
Video Virtual Try-on with Conditional Diffusion Transformer InpainterDiffusion models → 3D, point cloud and scene generation
Agent-RewardBench: Towards a Unified Benchmark for Reward Modeling across Perception, Planning, and Safety in Real-World Multimodal AgentsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +2
World-aware Planning Narratives Enhance Large Vision-Language Model PlannerLLM reasoning → Spatial and embodied reasoning +1
Unveiling Causal Reasoning in Large Language Models: Reality or Mirage?LLM reasoning → Causal reasoning in LLMs
T3: Multi-level Tree-based Automatic Program Repair with Large Language ModelsLLM reasoning → Reasoning for agents, planning and code
BitMark: Watermarking Bitwise Autoregressive Image Generative ModelsLLM safety and security → Watermarking generated content and models
GroundFlow: A Plug-in Module for Temporal Reasoning on 3D Point Cloud Sequential GroundingLLM reasoning → Spatial and embodied reasoning
Task-Aware KV Compression For Cost-Effective Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Cosmos: Compressed and Smooth Latent Space for Text Diffusion ModelingDiffusion models → Discrete and text diffusion
Diverse Mini-Batch Selection in Reinforcement Learning for Efficient Chemical Exploration in de novo Drug DesignDeep reinforcement learning methods → RL for energy, networks and operations
Geometry and Perception Guided Gaussians for Multiview-consistent 3D Generation from a Single ImageDiffusion models → 3D, point cloud and scene generation
Multi-Sample Prompting and Actor-Critic Prompt Optimization for Diverse Synthetic Data GenerationSynthetic data and self-improvement for LLMs → LLM synthetic data generation +1
Curriculum-Adapted Robust Reinforcement Learning for UAV Deconfliction in Adversarial EnvironmentsDeep reinforcement learning methods → RL for robot navigation and UAV control
Meta Policy Switching for Secure UAV Deconfliction in Adversarial AirspaceDeep reinforcement learning methods → RL for robot navigation and UAV control
GoIRL: Graph-Oriented Inverse Reinforcement Learning for Multimodal Trajectory PredictionOffline RL and imitation learning from data → Imitation and offline RL for driving +1
Doloris: Dual Conditional Diffusion Implicit Bridges with Sparsity Masking Strategy for Unpaired Single-Cell Perturbation EstimationDiffusion models → Flow matching, time series and science
Learning to Skip the Middle Layers of TransformersMixture-of-experts → Expert routing strategies
ComRAG: Retrieval-Augmented Generation with Dynamic Vector Stores for Real-time Community Question Answering in IndustryRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Chain-of-Thought Enhanced Shallow Transformers for Wireless Symbol DetectionLLM reasoning → Prompting and chain-of-thought reasoning
Enhancing LLM Tool Use with High-quality Instruction Data from Knowledge GraphLLM agents and tool use → Tool learning and function calling
Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse ConditionsOffline RL and imitation learning from data → Robot manipulation imitation learning
Improving Diffusion-Based Image Editing Faithfulness via Guidance and SchedulingDiffusion models → Image generation, text-to-image and editing
Efficient Skill Discovery via Regret-Aware OptimizationDeep reinforcement learning methods → Deep RL algorithms and representations
Boosting Domain Generalized and Adaptive Detection with Diffusion Models: Fitness, Generalization, and TransferabilityDiffusion models → Image generation, text-to-image and editing
SEAL: Vision-Language Model-Based Safe End-to-End Cooperative Autonomous Driving with Adaptive Long-Tail ModelingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement LearningDeep reinforcement learning methods → Deep RL algorithms and representations
RL-Selector: Reinforcement Learning-Guided Data Selection via Redundancy AssessmentDeep reinforcement learning methods → Deep RL algorithms and representations
Little by Little: Continual Learning via Incremental Mixture of Rank-1 Associative Memory ExpertsMixture-of-experts → Mixture of LoRA experts
Multimodal Prompt Alignment for Facial Expression RecognitionMultimodal LLMs / vision-language models → Multimodal emotion understanding
Distilling Normalizing FlowsKnowledge distillation → General knowledge distillation methods
MR-COSMO: Visual-Text Memory Recall and Direct CrOSs-MOdal Alignment Method for Query-Driven 3D SegmentationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
SharpZO: Hybrid Sharpness-Aware Vision Language Model Prompt Tuning via Forward-Only PassesMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
EVA: Mixture-of-Experts Semantic Variant Alignment for Compositional Zero-Shot LearningMixture-of-experts → Multimodal MoE models
Response Quality Assessment for Retrieval-Augmented Generation via Conditional Conformal FactualityRetrieval-augmented generation → RAG robustness and adaptive retrieval
DFVEdit: Conditional Delta Flow Vector for Zero-shot Video EditingDiffusion models → Video generation
Parallels Between VLA Model Post-Training and Human Motor Learning: Progress, Challenges, and TrendsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Evidence-based diagnostic reasoning with multi-agent copilot for human pathologyLLM reasoning → Clinical and biomedical applications +1
EraRAG: Efficient and Incremental Retrieval Augmented Generation for Growing CorporaRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Antibody Design and Optimization with Multi-scale Equivariant Graph Diffusion Models for Accurate Complex Antigen BindingDiffusion models → Flow matching, time series and science
Beyond Reactive Safety: Risk-Aware LLM Alignment via Long-Horizon SimulationLLM agents and tool use → Agent threats, guardrails and governance +1
Revisiting the Past: Data Unlearning with Model State HistoryLLM safety and security → Machine unlearning
Quantum Reinforcement Learning Trading Agent for Sector Rotation in the Taiwan Stock MarketDeep reinforcement learning methods → Adversarial robustness and RL in finance

All 331