Papers

12094 papers

ICML2025

PAK-UCB Contextual Bandit: An Online Learning Approach to Prompt-Aware Selection of Generative Models and LLMs

Summary pending...

Online LearningLarge Language ModelsEvaluation and Selection of Generative Models
ICML2025

CogMath: Assessing LLMs' Authentic Mathematical Ability from a Human Cognitive Perspective

Summary pending...

Large Language ModelsHuman CognitionEvaluation
ICML2025

Energy-Based Preference Model Offers Better Offline Alignment than the Bradley-Terry Preference Model

Summary pending...

preference modeloffline alignment
ICML2025

Taming Rectified Flow for Inversion and Editing

Summary pending...

Diffusion; Image Editing; Video Editing
ICML2025

Global Context-aware Representation Learning for Spatially Resolved Transcriptomics

Summary pending...

Spatially Resolved TranscriptomicsSelf-Supervised Learning
ICML2025

Scaling Laws for Pre-training Agents and World Models

Summary pending...

world modelingimitation learningscaling laws
ICML2025

Do Not Mimic My Voice : Speaker Identity Unlearning for Zero-Shot Text-to-Speech

Summary pending...

machine unlearningzero-shot ttsvoice privacy
ICML2025

Earley-Driven Dynamic Pruning for Efficient Structured Decoding

Summary pending...

constrained decodingstructured generationLLM inference
ICML2025

Learning Imbalanced Data with Beneficial Label Noise

Summary pending...

Imbalanced learningbeneficial label noiseclassification accuracy
ICML2025

OrthoRank: Token Selection via Sink Token Orthogonality for Efficient LLM inference

Summary pending...

large language modelattention sinkefficiency
ICML2025

Foundation Model Insights and a Multi-Model Approach for Superior Fine-Grained One-shot Subset Selection

Summary pending...

one-shot subset selectionfoundation modelsdata-efficient learning
ICML2025

Smooth Interpolation for Improved Discrete Graph Generative Models

Summary pending...

Graph Generation; Deep Generative Models; Bayesian Flow Networks;
ICML2025

CAD-Editor: A Locate-then-Infill Framework with Automated Training Data Synthesis for Text-Based CAD Editing

Summary pending...

Computer Aided DesignGenerative ModelsText-based Editing
ICML2025

BSLoRA: Enhancing the Parameter Efficiency of LoRA with Intra-Layer and Inter-Layer Sharing

Summary pending...

Parameter-efficient fine-tuningparameter-sharing
ICML2025

Multi-objective Linear Reinforcement Learning with Lexicographic Rewards

Summary pending...

Reinforcement learningmulti-objectivelinear model
ICML2025

FourierMamba: Fourier Learning Integration with State Space Models for Image Deraining

Summary pending...

FourierderainingMamba
ICML2025

End-to-End Learning Framework for Solving Non-Markovian Optimal Control

Summary pending...

Fractional-order dynamical systemOptimal control
ICML2025

ReFrame: Layer Caching for Accelerated Inference in Real-Time Rendering

Summary pending...

Computer GraphicsNeural Network InferenceTraining-Free Optimization
ICML2025

Code-Generated Graph Representations Using Multiple LLM Agents for Material Properties Prediction

Summary pending...

Material structure representationMaterial property predictionMulti-agent systems
ICML2025

Reasoning-as-Logic-Units: Scaling Test-Time Reasoning in Large Language Models Through Logic Unit Alignment

Summary pending...

large language modelsLLMsreasoning