Papers

New on arXiv

Sun 20 Apr 2025

138 papers

Perfect diffusion is mathsfTC0 -- Bad diffusion is Turing-completeDiffusion models → Discrete and text diffusion
Deep Reinforcement Learning for Investor-Specific Portfolio Optimization: A Volatility-Guided Asset Selection ApproachDeep reinforcement learning methods → Adversarial robustness and RL in finance
Breaking the Prompt Wall (I): A Real-World Case Study of Attacking ChatGPT via Lightweight Prompt InjectionLLM safety and security → Prompt injection and agent security
Diffusion-Driven Inertial Generated Data for Smartphone Location ClassificationDiffusion models → Video generation
PolicyEvol-Agent: Evolving Policy via Environment Perception and Self-Awareness with Theory of MindLLM agents and tool use → Reinforcement learning for LLM agents
Knowledge Distillation and Dataset Distillation of Large Language Models: Emerging Trends, Challenges, and Future DirectionsKnowledge distillation → Knowledge distillation for LLMs
PROMPTEVALS: A Dataset of Assertions and Guardrails for Custom Production Large Language Model PipelinesLLM safety and security → Safety alignment, refusal and guardrails
Reinforcement Learning from Multi-level and Episodic Human FeedbackLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Pairwise or Pointwise? Evaluating Feedback Protocols for Bias in LLM-Based EvaluationLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Exposing the Copycat Problem of Imitation-based Planner: A Novel Closed-Loop Simulator, Causal Benchmark and Joint IL-RL BaselineOffline RL and imitation learning from data → Imitation and offline RL for driving +1
Learning Critically: Selective Self Distillation in Federated Learning on Non-IID DataKnowledge distillation → Federated, robust and private distillation
OmniV-Med: Scaling Medical Vision-Language Model for Universal Visual UnderstandingMultimodal LLMs / vision-language models → Medical vision-language models
An LLM-enabled Multi-Agent Autonomous Mechatronics Design FrameworkLLM agents and tool use → General LLM agent frameworks and benchmarks
Trans-Zero: Self-Play Incentivizes Large Language Models for Multilingual Translation Without Parallel DataSynthetic data and self-improvement for LLMs → LLM synthetic data generation
Generative Multimodal Pretraining with Discrete Diffusion Timestep TokensMultimodal LLMs / vision-language models → Image editing and quality assessment +1
LeetCodeDataset: A Temporal Dataset for Robust Evaluation and Efficient Training of Code LLMsLLM reasoning → Reasoning for agents, planning and code
A Framework for Benchmarking and Aligning Task-Planning Safety in LLM-Based Embodied AgentsLLM agents and tool use → Agent threats, guardrails and governance
Surrogate Fitness Metrics for Interpretable Reinforcement LearningDeep reinforcement learning methods → Interpretable and logic-guided RL
Relation-R1: Progressively Cognitive Chain-of-Thought Guided Reinforcement Learning for Unified Relation ComprehensionLLM reasoning → Multimodal reasoning +1
NVSMask3D: Hard Visual Prompting with Camera Pose Interpolation for 3D Open Vocabulary Instance SegmentationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Towards Optimal Circuit Generation: Multi-Agent Collaboration Meets Collective IntelligenceLLM agents and tool use → Software engineering coding agents
VM-BHINet:Vision Mamba Bimanual Hand Interaction Network for 3D Interacting Hand Mesh Recovery From a Single RGB ImageState-space sequence models (S4, Mamba) → Video, motion and generative Mamba models
a1: Steep Test-time Scaling Law via Environment Augmented GenerationLLM reasoning → Test-time compute and efficient thinking
ReasoningV: Efficient Verilog Code Generation with Adaptive Hybrid Reasoning ModelLLM reasoning → Reasoning for agents, planning and code
Enhancing LLM-based Quantum Code Generation with Multi-Agent Optimization and Quantum Error CorrectionLLM agents and tool use → Medical and scientific domain agents
REDEditing: Relationship-Driven Precise Backdoor Poisoning on Text-to-Image Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
BookWorld: From Novels to Interactive Agent Societies for Creative Story GenerationLLM agents and tool use → Multimodal and embodied agents
FlowLoss: Dynamic Flow-Conditioned Loss Strategy for Video Diffusion ModelsDiffusion models → Video generation
SUDO: Enhancing Text-to-Image Diffusion Models with Self-Supervised Direct Preference OptimizationDiffusion models → Diffusion and flow policies for robotics +1
Causality for Natural Language ProcessingLLM reasoning → Causal reasoning in LLMs
Are Vision LLMs Road-Ready? A Comprehensive Benchmark for Safety-Critical Driving Video UnderstandingEnd-to-end autonomous driving → VLMs for autonomous driving
Biased by Design: Leveraging AI Biases to Enhance Critical Thinking of News ReadersLLM reasoning → Reasoning benchmarks and evaluation
Meta-Thinking in LLMs via Multi-Agent Reinforcement Learning: A SurveyLLM reasoning → Reasoning for agents, planning and code +1
DreamID: High-Fidelity and Fast diffusion-based Face Swapping via Triplet ID Group LearningDiffusion models → Image generation, text-to-image and editing
Less is More: Adaptive Coverage for Synthetic Training DataSynthetic data and self-improvement for LLMs → LLM synthetic data generation
FinSage: A Multi-aspect RAG System for Financial Filings Question AnsweringRetrieval-augmented generation → Legal and financial RAG
FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation SteeringMechanistic interpretability → Probing deception, bias and beliefs
ExFace: Expressive Facial Control for Humanoid Robots with Diffusion Transformers and Bootstrap TrainingDiffusion models → Audio, speech and music
LGD: Leveraging Generative Descriptions for Zero-Shot Referring Image SegmentationMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
CoLoTa: A Dataset for Entity-based Commonsense Reasoning over Long-Tail KnowledgeLLM reasoning → Reasoning benchmarks and evaluation
ParaPO: Aligning Language Models to Reduce Verbatim Reproduction of Pre-training DataLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Causal Disentanglement for Robust Long-tail Medical Image GenerationDiffusion models → Medical imaging and segmentation
LoRe: Personalizing LLMs via Low-Rank Reward ModelingLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Diffusion-based supervised learning of generative models for efficient sampling of multimodal distributions
Efficient Knowledge Transfer in Multi-Task Learning through Task-Adaptive Low-Rank Representation
Don't Retrieve, Generate: Prompting LLMs for Synthetic Training Data in Dense Retrieval
Deep Learning with Pretrained 'Internal World' Layers: A Gemma 3-Based Modular Architecture for Wildfire Prediction
Enhancing DR Classification with Swin Transformer and Shifted Window Attention
M-TabNet: A Multi-Encoder Transformer Model for Predicting Neonatal Birth Weight from Multimodal Data
Disentangling Linguistic Features with Dimension-Wise Analysis of Vector Embeddings
A Combinatorial Theory of Dropout: Subnetworks, Graph Geometry, and Generalization
SWE-Synth: Synthesizing Verifiable Bug-Fix Data to Enable Large Language Models in Resolving Real-World Bugs
Advancing Video Anomaly Detection: A Bi-Directional Hybrid Framework for Enhanced Single- and Multi-Task Approaches
AI for the Open-World: the Learning Principles
On the Tunability of Random Survival Forests Model for Predictive Maintenance
AltGDmin: Alternating GD and Minimization for Partly-Decoupled (Federated) Optimization
A Modularized Design Approach for GelSight Family of Vision-based Tactile Sensors
SuperCL: Superpixel Guided Contrastive Learning for Medical Image Segmentation Pre-training
ChronoRoot 2.0: An Open AI-Powered Platform for 2D Temporal Plant Phenotyping
Geometric Learning Dynamics
Semi-parametric Memory Consolidation: Towards Brain-like Deep Continual Learning
Video QoE Metrics from Encrypted Traffic: Application-agnostic Methodology
TAPIP3D: Tracking Any Point in Persistent 3D Geometry
Med-2D SegNet: A Light Weight Deep Neural Network for Medical 2D Image Segmentation
Time Frequency Analysis of EMG Signal for Gesture Recognition using Fine grained Features
FLAME: A New Dataset on FLemish Accounts of Momentary Experiences
AI with Emotions: Exploring Emotional Expressions in Large Language Models
Can We Ignore Labels In Out of Distribution Detection?
Connecting Parameter Magnitudes and Hessian Eigenspaces at Scale using Sketched Methods
IXGS-Intraoperative 3D Reconstruction from Sparse, Arbitrarily Posed Real X-rays
Quantitative Clustering in Mean-Field Transformer Models
Reveal-or-Obscure: A Differentially Private Sampling Algorithm for Discrete Distributions
Video-MMLU: A Massive Multi-Discipline Lecture Understanding Benchmark
FarsEval-PKBETS: A new diverse benchmark for evaluating Persian large language models
Seurat: From Moving Points to Depth
Uncovering Issues in the Radio Access Network by Looking at the Neighbors
Evaluating Temporal Plasticity in Foundation Time Series Models for Incremental Fine-tuning
Efficient Split Federated Learning for Large Language Models over Communication Networks
DMPCN: Dynamic Modulated Predictive Coding Network with Hybrid Feedback Representations
Frequency-domain Learning with Kernel Prior for Blind Image Deblurring
Mitigating Parameter Interference in Model Merging via Sharpness-Aware Fine-Tuning
EmoSEM: Segment and Explain Emotion Stimuli in Visual Art
A Case Study Exploring the Current Landscape of Synthetic Medical Record Generation with Commercial LLMs
Risk Assessment Framework for Code LLMs via Leveraging Internal States
When Search Teaches Style: Causal Internalization of Tactical Priors in AlphaZero
Latent Representations for Visual Proprioception in Inexpensive Robots
Harnessing Generative LLMs for Enhanced Financial Event Entity Extraction Performance
Automatic Text Summarization (ATS) for Research Documents in Sorani Kurdish
GENE-FL: Gene-Driven Parameter-Efficient Dynamic Federated Learning
MSAD-Net: Multiscale and Spatial Attention-based Dense Network for Lung Cancer Classification
Consensus in Motion: A Case of Dynamic Rationality of Sequential Learning in Probability Aggregation
Talk is Not Always Cheap: Promoting Wireless Sensing Models with Text Prompts
A Hierarchical Framework for Measuring Scientific Paper Innovation via Large Language Models
Translation Analytics for Freelancers: I. Introduction, Data Preparation, Baseline Evaluations
Imputation-free Learning of Tabular Data with Missing Values using Incremental Feature Partitions in Transformer
MP-Mat: A 3D-and-Instance-Aware Human Matting and Editing Framework with Multiplane Representation
Generalized Derangetropy Functionals for Modeling Cyclical Information Flow
UFO2: The Desktop AgentOS
K2MUSE: A human lower-limb multimodal walking dataset spanning task and acquisition variability for rehabilitation robotics
NTIRE 2025 Challenge on Real-World Face Restoration: Methods and Results

All 138