Papers

12094 papers

ICML2025

Towards Cost-Effective Reward Guided Text Generation

Summary pending...

LLMRLHFAlignment
ICML2025

Efficient Optimization with Orthogonality Constraint: a Randomized Riemannian Submanifold Method

Summary pending...

Riemannian manifoldStiefel manifoldEfficient optimization
ICML2025

Regress, Don't Guess: A Regression-like Loss on Number Tokens for Language Models

Summary pending...

language modelsmathematical reasoningarithmetics
ICML2025

Can We Predict Performance of Large Models across Vision-Language Tasks?

Summary pending...

Large Vision-Language Models (LVLMs)BenchmarkingProbabilistic Matrix Factorization (PMF)
ICML2025

Towards a Mechanistic Explanation of Diffusion Model Generalization

Summary pending...

Diffusion ModelsGeneralizationDiffusion
ICML2025

Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond

Summary pending...

Large Language ModelMachine UnlearningRobustness
ICML2025

Learn from Downstream and Be Yourself in Multimodal Large Language Models Fine-Tuning

Summary pending...

Multimodal Large Language ModelsFine-TuningCatastrophic Forgetting
ICML2025

Improving the Variance of Differentially Private Randomized Experiments through Clustering

Summary pending...

causal inferencedifferential privacyclustering
ICML2025

MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents

Summary pending...

Indirect prompt injectionAgent system for tool useLarge langugae models
ICML2025

Automated Hypothesis Validation with Agentic Sequential Falsifications

Summary pending...

LLM agenthypothesis testingsequential decision making
ICML2025

Leveraging Offline Data in Linear Latent Contextual Bandits

Summary pending...

banditslatent banditshybrid RL
ICML2025

Optimizing Robustness and Accuracy in Mixture of Experts: A Dual-Model Approach

Summary pending...

Mixture of ExpertsAdversarial DefenseAdversarial Robustness
ICML2025

Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks' Internal Representations

Summary pending...

Concept based Explainable AITCAVVision-Language Models
ICML2025

An Asymptotically Optimal Approximation Algorithm for Multiobjective Submodular Maximization at Scale

Summary pending...

submodular maximizationmultiobjectivemax-min fairness
ICML2025

Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

Summary pending...

evolutionary algorithmmodel adaptationswarm intelligence
ICML2025

OrcaLoca: An LLM Agent Framework for Software Issue Localization

Summary pending...

LLMLLM AgentSoftware Engineering
ICML2025

WOMD-Reasoning: A Large-Scale Dataset for Interaction Reasoning in Driving

Summary pending...

Language Q&A DatasetAutonomous DrivingInteraction Reasoning
ICML2025

SEAD: Unsupervised Ensemble of Streaming Anomaly Detectors

Summary pending...

anomaly detectiononline learningstreaming
ICML2025

Concurrent Reinforcement Learning with Aggregated States via Randomized Least Squares Value Iteration

Summary pending...

worst-case regret boundRL theoryrandomized least squares value iteration
ICML2025

Contrastive Localized Language-Image Pre-Training

Summary pending...

CLIPMLLMFoundation Models