Learning bounds for doubly-robust covariate shift adaptation

Jeonghwan Lee, Cong Ma

Distribution shift between the training domain and the test domain poses a key challenge for modern machine learning. An extensively studied instance is the \emph{covariate shift},…