Optimal transport for conditional domain matching and label shift

We address the problem of unsupervised domain adaptation under the setting of generalized target shift (joint class-conditional and label shifts). For this framework, we theoretically show that, for good generalization, it is necessary to learn a latent representation in which both marginals and class-conditional distributions are aligned across domains. For this sake, we propose a learning problem that minimizes importance weighted loss in the source domain and a Wasserstein distance between weighted marginals. For a proper weighting, we provide an estimator of target label proportion by blending mixture estimation and optimal matching by optimal transport. This estimation comes with theoretical guarantees of correctness under mild assumptions. Our experimental results show that our method performs better on average than competitors across a range domain adaptation problems including digits,VisDA and Office. Code for this paper is available at https://github.com/arakotom/mars_domain_adaptation.

Automated summary

This study addresses the problem of unsupervised domain adaptation under generalized target shift, showing the importance of aligning marginals and class-conditional distributions across domains for good generalization. The proposed learning framework minimizes importance weighted loss in the source domain and utilizes a Wasserstein distance between weighted marginals, with an estimator of target label proportion blending mixture estimation and optimal matching by optimal transport. Experimental results demonstrate the method outperforms competitors on various domain adaptation problems.