Optimal Transport Based Domain Adaptation

In this project, we focus on a large-scale optimal transportation model to perform the alignment of the representations in the source and target domains.

Recently, adapting or transferring knowledge across different datasets or domains has gained increasing interests. Different learning methods include domain adaptation, transfer learning, meta learning and few/zero-show learning. Among the many strategies proposed to adapt a domain to another, finding a common representation has shown excellent properties: by finding a common representation for both domains, a single classifier can be effective in both and use labelled samples from the source domain to predict the unlabelled samples of the target domain. In this project, we focus on a large-scale optimal transportation model to perform the alignment of the representations in the source and target domains. First, we learn an optimal transport (OT) plan. To that end, we use a stochastic dual approach of regularized OT, which enables OT scale to large datasets. Second, we estimate a \textit{Monge map} as a deep neural network learned by approximating the barycentric projection of the previously-obtained OT plan. This parameterization allows generalization of the mapping outside the support of the input measure. The source code for this project can be obtained by this github repo: \url{https://github.com/yihhhh/OT-for-Domain-Adaptation}.