Domain Adaptation with Conditional Distribution Matching and Generalized Label Shift

Adversarial learning has demonstrated good performance in the unsupervised domain adaptation setting, by learning domain-invariant representations. However, recent work has shown limitations of this approach when label distributions differ between the source and target domains. In this paper, we propose a new assumption, \textit{generalized label shift} ( \glsa), to improve robustness against mismatched label distributions. Under \glsa, we provide theoretical guarantees on the transfer performance of any classifier. We also devise necessary and sufficient conditions for \glsa to hold, by using an estimation of the relative class weights between domains and an appropriate reweighting of samples.