In-Context Learning in Large Language Models Learns Label Relationships but Is Not Conventional Learning

Three Towers: Flexible Contrastive Learning with Pretrained Image Models

Active Surrogate Estimators: An Active Learning Approach to Label-Efficient Model Evaluation

Self-Attention Between Datapoints: Going Beyond Individual Input-Output Pairs in Deep Learning

Active Testing: Sample-Efficient Model Evaluation

Structured object-aware physics prediction for video modeling and planning