Learning Unforeseen Robustness from Out-of-distribution Data Using Equivariant Domain Translator

Sicheng Zhu, Bang An, Furong Huang, Sanghyun Hong
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:42915-42937, 2023.

Abstract

Current approaches for training robust models are typically tailored to scenarios where data variations are accessible in the training set. While shown effective in achieving robustness to these foreseen variations, these approaches are ineffective in learning unforeseen robustness, i.e., robustness to data variations without known characterization or training examples reflecting them. In this work, we learn unforeseen robustness by harnessing the variations in the abundant out-of-distribution data. To overcome the main challenge of using such data, the domain gap, we use a domain translator to bridge it and bound the unforeseen robustness on the target distribution. As implied by our analysis, we propose a two-step algorithm that first trains an equivariant domain translator to map out-of-distribution data to the target distribution while preserving the considered variation, and then regularizes a model’s output consistency on the domain-translated data to improve its robustness. We empirically show the effectiveness of our approach in improving unforeseen and foreseen robustness compared to existing approaches. Additionally, we show that training the equivariant domain translator serves as an effective criterion for source data selection.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-zhu23a, title = {Learning Unforeseen Robustness from Out-of-distribution Data Using Equivariant Domain Translator}, author = {Zhu, Sicheng and An, Bang and Huang, Furong and Hong, Sanghyun}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {42915--42937}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/zhu23a/zhu23a.pdf}, url = {https://proceedings.mlr.press/v202/zhu23a.html}, abstract = {Current approaches for training robust models are typically tailored to scenarios where data variations are accessible in the training set. While shown effective in achieving robustness to these foreseen variations, these approaches are ineffective in learning unforeseen robustness, i.e., robustness to data variations without known characterization or training examples reflecting them. In this work, we learn unforeseen robustness by harnessing the variations in the abundant out-of-distribution data. To overcome the main challenge of using such data, the domain gap, we use a domain translator to bridge it and bound the unforeseen robustness on the target distribution. As implied by our analysis, we propose a two-step algorithm that first trains an equivariant domain translator to map out-of-distribution data to the target distribution while preserving the considered variation, and then regularizes a model’s output consistency on the domain-translated data to improve its robustness. We empirically show the effectiveness of our approach in improving unforeseen and foreseen robustness compared to existing approaches. Additionally, we show that training the equivariant domain translator serves as an effective criterion for source data selection.} }
Endnote
%0 Conference Paper %T Learning Unforeseen Robustness from Out-of-distribution Data Using Equivariant Domain Translator %A Sicheng Zhu %A Bang An %A Furong Huang %A Sanghyun Hong %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-zhu23a %I PMLR %P 42915--42937 %U https://proceedings.mlr.press/v202/zhu23a.html %V 202 %X Current approaches for training robust models are typically tailored to scenarios where data variations are accessible in the training set. While shown effective in achieving robustness to these foreseen variations, these approaches are ineffective in learning unforeseen robustness, i.e., robustness to data variations without known characterization or training examples reflecting them. In this work, we learn unforeseen robustness by harnessing the variations in the abundant out-of-distribution data. To overcome the main challenge of using such data, the domain gap, we use a domain translator to bridge it and bound the unforeseen robustness on the target distribution. As implied by our analysis, we propose a two-step algorithm that first trains an equivariant domain translator to map out-of-distribution data to the target distribution while preserving the considered variation, and then regularizes a model’s output consistency on the domain-translated data to improve its robustness. We empirically show the effectiveness of our approach in improving unforeseen and foreseen robustness compared to existing approaches. Additionally, we show that training the equivariant domain translator serves as an effective criterion for source data selection.
APA
Zhu, S., An, B., Huang, F. & Hong, S.. (2023). Learning Unforeseen Robustness from Out-of-distribution Data Using Equivariant Domain Translator. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:42915-42937 Available from https://proceedings.mlr.press/v202/zhu23a.html.

Related Material