Papers

12094 papers

ICML2025

Signed Laplacians for Constrained Graph Clustering

Summary pending...

signed Laplaciansconstrained clusteringCheeger inequality
ICML2025

System-Aware Unlearning Algorithms: Use Lesser, Forget Faster

Summary pending...

machine unlearningsample compressionselective sampling
ICML2025

From Logits to Hierarchies: Hierarchical Clustering made Simple

Summary pending...

Hierarchical ClusteringClusteringInterpretability
ICML2025

Aligning Multimodal Representations through an Information Bottleneck

Summary pending...

multimodal representation learningrepresentational alignmentmodality gap
ICML2025

Metadata Conditioning Accelerates Language Model Pre-training

Summary pending...

language modelspre-trainingmetadata conditioning
ICML2025

Online Sparsification of Bipartite-Like Clusters in Graphs

Summary pending...

bipartite-like componentssparsificationonline algorithms
ICML2025

Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-off

Summary pending...

Federated LearningDifferential PrivacyUtility-Privacy Trade-off
ICML2025

The Devil Is in the Details: Tackling Unimodal Spurious Correlations for Generalizable Multimodal Reward Models

Summary pending...

Multimodal Reward ModelGeneralizationSpurious Correlations
ICML2025

Upcycling Text-to-Image Diffusion Models for Multi-Task Capabilities

Summary pending...

multi-task learningupcyclingdiffusion models
ICML2025

Flopping for FLOPs: Leveraging Equivariance for Computational Efficiency

Summary pending...

Equivarianceefficiencyscalable
ICML2025

Provably Cost-Sensitive Adversarial Defense via Randomized Smoothing

Summary pending...

Certificationcost-sensitive learning
ICML2025

MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking

Summary pending...

AI AlignmentMyopic OptimizationReinforcement Learning
ICML2025

Protein Structure Tokenization: Benchmarking and New Recipe

Summary pending...

protein structure tokenizationbenchmarkingVQ-VAE
ICML2025

LAuReL: Learned Augmented Residual Layer

Summary pending...

residual connectionskip connectionresidual layer
ICML2025

Towards Theoretical Understanding of Sequential Decision Making with Preference Feedback

Summary pending...

preference-basedsequential decision makingpreference feedback
ICML2025

Square$\chi$PO: Differentially Private and Robust $\chi^2$-Preference Optimization in Offline Direct Alignment

Summary pending...

Offline alignmentgeneral function approximationsdifferential privacy
ICML2025

LIVS: A Pluralistic Alignment Dataset for Inclusive Public Spaces

Summary pending...

Pluralistic AlignmentText-to-Image DiffusionIntersectionality
ICML2025

The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models Via Visual Information Steering

Summary pending...

Large Vision-Language ModelHallucination
ICML2025

MathConstruct: Challenging LLM Reasoning with Constructive Proofs

Summary pending...

LLMmathevaluation
ICML2025

The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence

Summary pending...

Refusal directionsLarge Language ModelActivation Steering