Multi-Source Domain Adaptation through Dataset Dictionary Learning in Wasserstein Space

TL;DR

This paper introduces a novel multi-source domain adaptation framework using dictionary learning and optimal transport, representing each domain as a Wasserstein barycenter of learned distributions to improve classification across domains.

Contribution

The paper proposes DaDiL, a new algorithm for learning domain atoms and barycentric coordinates, enabling effective MSDA through Wasserstein space representations.

Findings

Achieved state-of-the-art results on three benchmarks

Improved classification performance by up to 7.71%

Demonstrated effective domain generalization via Wasserstein barycenter interpolations

Abstract

This paper seeks to solve Multi-Source Domain Adaptation (MSDA), which aims to mitigate data distribution shifts when transferring knowledge from multiple labeled source domains to an unlabeled target domain. We propose a novel MSDA framework based on dictionary learning and optimal transport. We interpret each domain in MSDA as an empirical distribution. As such, we express each domain as a Wasserstein barycenter of dictionary atoms, which are empirical distributions. We propose a novel algorithm, DaDiL, for learning via mini-batches: (i) atom distributions; (ii) a matrix of barycentric coordinates. Based on our dictionary, we propose two novel methods for MSDA: DaDil-R, based on the reconstruction of labeled samples in the target domain, and DaDiL-E, based on the ensembling of classifiers learned on atom distributions. We evaluate our methods in 3 benchmarks: Caltech-Office, Office 31, and CRWU, where we improved previous state-of-the-art by 3.15%, 2.29%, and 7.71% in classification performance. Finally, we show that interpolations in the Wasserstein hull of learned atoms provide data that can generalize to the target domain.