Adam Tauman Kalai

Learn to Expect the Unexpected: Probably Approximately Correct Domain Generalization

2020 · arXiv (Cornell University)

Domain generalization is the problem of machine learning when the training data and the test data come from different data domains. We present a simple theoretical model of learning to generalize across domains in which …

Identifying unpredictable test examples with worst-case guarantees

2020

Often times, whether it be for adversarial or natural reasons, the distributions of test and training data differ. We give an algorithm that, given sets of training and test examples, identifies regions of test examples …