Make the Most of Your Data: Changing the Training Data Distribution to Improve In-distribution Generalization Performance

Open in new window