Multi-source domain adaptation via weighted joint distributions optimal transport

Rosanna Turrisi, Rémi Flamary, Alain Rakotomamonjy, Massimiliano Pontil
Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence, PMLR 180:1970-1980, 2022.

Abstract

This work addresses the problem of domain adaptation on an unlabeled target dataset using knowledge from multiple labelled source datasets. Most current approaches tackle this problem by searching for an embedding that is invariant across source and target domains, which corresponds to searching for a universal classifier that works well on all domains. In this paper, we address this problem from a new perspective: instead of crushing diversity of the source distributions, we exploit it to adapt better to the target distribution. Our method, named Multi-Source Domain Adaptation via Weighted Joint Distribution Optimal Transport (MSDA-WJDOT), aims at finding simultaneously an Optimal Transport-based alignment between the source and target distributions and a re-weighting of the sources distributions. We discuss the theoret- ical aspects of the method and propose a conceptually simple algorithm. Numerical experiments indicate that the proposed method achieves state-of- the-art performance on simulated and real datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v180-turrisi22a, title = {Multi-source domain adaptation via weighted joint distributions optimal transport}, author = {Turrisi, Rosanna and Flamary, R\'emi and Rakotomamonjy, Alain and Pontil, Massimiliano}, booktitle = {Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence}, pages = {1970--1980}, year = {2022}, editor = {Cussens, James and Zhang, Kun}, volume = {180}, series = {Proceedings of Machine Learning Research}, month = {01--05 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v180/turrisi22a/turrisi22a.pdf}, url = {https://proceedings.mlr.press/v180/turrisi22a.html}, abstract = {This work addresses the problem of domain adaptation on an unlabeled target dataset using knowledge from multiple labelled source datasets. Most current approaches tackle this problem by searching for an embedding that is invariant across source and target domains, which corresponds to searching for a universal classifier that works well on all domains. In this paper, we address this problem from a new perspective: instead of crushing diversity of the source distributions, we exploit it to adapt better to the target distribution. Our method, named Multi-Source Domain Adaptation via Weighted Joint Distribution Optimal Transport (MSDA-WJDOT), aims at finding simultaneously an Optimal Transport-based alignment between the source and target distributions and a re-weighting of the sources distributions. We discuss the theoret- ical aspects of the method and propose a conceptually simple algorithm. Numerical experiments indicate that the proposed method achieves state-of- the-art performance on simulated and real datasets.} }
Endnote
%0 Conference Paper %T Multi-source domain adaptation via weighted joint distributions optimal transport %A Rosanna Turrisi %A Rémi Flamary %A Alain Rakotomamonjy %A Massimiliano Pontil %B Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2022 %E James Cussens %E Kun Zhang %F pmlr-v180-turrisi22a %I PMLR %P 1970--1980 %U https://proceedings.mlr.press/v180/turrisi22a.html %V 180 %X This work addresses the problem of domain adaptation on an unlabeled target dataset using knowledge from multiple labelled source datasets. Most current approaches tackle this problem by searching for an embedding that is invariant across source and target domains, which corresponds to searching for a universal classifier that works well on all domains. In this paper, we address this problem from a new perspective: instead of crushing diversity of the source distributions, we exploit it to adapt better to the target distribution. Our method, named Multi-Source Domain Adaptation via Weighted Joint Distribution Optimal Transport (MSDA-WJDOT), aims at finding simultaneously an Optimal Transport-based alignment between the source and target distributions and a re-weighting of the sources distributions. We discuss the theoret- ical aspects of the method and propose a conceptually simple algorithm. Numerical experiments indicate that the proposed method achieves state-of- the-art performance on simulated and real datasets.
APA
Turrisi, R., Flamary, R., Rakotomamonjy, A. & Pontil, M.. (2022). Multi-source domain adaptation via weighted joint distributions optimal transport. Proceedings of the Thirty-Eighth Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 180:1970-1980 Available from https://proceedings.mlr.press/v180/turrisi22a.html.

Related Material