Rethinking Importance Weighting for Deep Learning under Distribution Shift

Meta Review

This paper works on importance weighting for deep learning under distribution shift. All reviewers agree that the manuscript is of high quality, that the ideas are novel and the results are significant.