Fetching the paper…

Rethinking Importance Weighting for Deep Learning under Distribution Shift · Around