A Minimalist Approach for Domain Adaptation with Optimal Transport

Arip Asadulaev, Vitaly Shutov, Alexander Korotin, Alexander Panfilov, Vladislava Kontsevaya, Andrey Filchenkov
Proceedings of The 2nd Conference on Lifelong Learning Agents, PMLR 232:1009-1024, 2023.

Abstract

We reveal an intriguing connection between adversarial attacks and cycle monotone maps, also known as optimal transport maps. Based on this finding, we developed a novel method named \textit{source fiction} for semi-supervised optimal transport-based domain adaptation. We conduct experiments on various datasets and show that our method can notably improve the performance of the optimal transport solvers in domain adaptation.

Cite this Paper

BibTeX
@InProceedings{pmlr-v232-asadulaev23a, title = {A Minimalist Approach for Domain Adaptation with Optimal Transport}, author = {Asadulaev, Arip and Shutov, Vitaly and Korotin, Alexander and Panfilov, Alexander and Kontsevaya, Vladislava and Filchenkov, Andrey}, booktitle = {Proceedings of The 2nd Conference on Lifelong Learning Agents}, pages = {1009--1024}, year = {2023}, editor = {Chandar, Sarath and Pascanu, Razvan and Sedghi, Hanie and Precup, Doina}, volume = {232}, series = {Proceedings of Machine Learning Research}, month = {22--25 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v232/asadulaev23a/asadulaev23a.pdf}, url = {https://proceedings.mlr.press/v232/asadulaev23a.html}, abstract = {We reveal an intriguing connection between adversarial attacks and cycle monotone maps, also known as optimal transport maps. Based on this finding, we developed a novel method named \textit{source fiction} for semi-supervised optimal transport-based domain adaptation. We conduct experiments on various datasets and show that our method can notably improve the performance of the optimal transport solvers in domain adaptation.} }
Endnote
%0 Conference Paper %T A Minimalist Approach for Domain Adaptation with Optimal Transport %A Arip Asadulaev %A Vitaly Shutov %A Alexander Korotin %A Alexander Panfilov %A Vladislava Kontsevaya %A Andrey Filchenkov %B Proceedings of The 2nd Conference on Lifelong Learning Agents %C Proceedings of Machine Learning Research %D 2023 %E Sarath Chandar %E Razvan Pascanu %E Hanie Sedghi %E Doina Precup %F pmlr-v232-asadulaev23a %I PMLR %P 1009--1024 %U https://proceedings.mlr.press/v232/asadulaev23a.html %V 232 %X We reveal an intriguing connection between adversarial attacks and cycle monotone maps, also known as optimal transport maps. Based on this finding, we developed a novel method named \textit{source fiction} for semi-supervised optimal transport-based domain adaptation. We conduct experiments on various datasets and show that our method can notably improve the performance of the optimal transport solvers in domain adaptation.
APA
Asadulaev, A., Shutov, V., Korotin, A., Panfilov, A., Kontsevaya, V. & Filchenkov, A.. (2023). A Minimalist Approach for Domain Adaptation with Optimal Transport. Proceedings of The 2nd Conference on Lifelong Learning Agents, in Proceedings of Machine Learning Research 232:1009-1024 Available from https://proceedings.mlr.press/v232/asadulaev23a.html.

Related Material