Towards Practical Non-Adversarial Distribution Matching

Ziyu Gong, Ben Usman, Han Zhao, David I Inouye
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, PMLR 238:4276-4284, 2024.

Abstract

Distribution matching can be used to learn invariant representations with applications in fairness and robustness. Most prior works resort to adversarial matching methods but the resulting minimax problems are unstable and challenging to optimize. Non-adversarial likelihood-based approaches either require model invertibility, impose constraints on the latent prior, or lack a generic framework for distribution matching. To overcome these limitations, we propose a non-adversarial VAE-based matching method that can be applied to any model pipeline. We develop a set of alignment upper bounds for distribution matching (including a noisy bound) that have VAE-like objectives but with a different perspective. We carefully compare our method to prior VAE-based matching approaches both theoretically and empirically. Finally, we demonstrate that our novel matching losses can replace adversarial losses in standard invariant representation learning pipelines without modifying the original architectures—thereby significantly broadening the applicability of non-adversarial matching methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v238-gong24b, title = { Towards Practical Non-Adversarial Distribution Matching }, author = {Gong, Ziyu and Usman, Ben and Zhao, Han and I Inouye, David}, booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, pages = {4276--4284}, year = {2024}, editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, volume = {238}, series = {Proceedings of Machine Learning Research}, month = {02--04 May}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v238/gong24b/gong24b.pdf}, url = {https://proceedings.mlr.press/v238/gong24b.html}, abstract = { Distribution matching can be used to learn invariant representations with applications in fairness and robustness. Most prior works resort to adversarial matching methods but the resulting minimax problems are unstable and challenging to optimize. Non-adversarial likelihood-based approaches either require model invertibility, impose constraints on the latent prior, or lack a generic framework for distribution matching. To overcome these limitations, we propose a non-adversarial VAE-based matching method that can be applied to any model pipeline. We develop a set of alignment upper bounds for distribution matching (including a noisy bound) that have VAE-like objectives but with a different perspective. We carefully compare our method to prior VAE-based matching approaches both theoretically and empirically. Finally, we demonstrate that our novel matching losses can replace adversarial losses in standard invariant representation learning pipelines without modifying the original architectures—thereby significantly broadening the applicability of non-adversarial matching methods. } }
Endnote
%0 Conference Paper %T Towards Practical Non-Adversarial Distribution Matching %A Ziyu Gong %A Ben Usman %A Han Zhao %A David I Inouye %B Proceedings of The 27th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2024 %E Sanjoy Dasgupta %E Stephan Mandt %E Yingzhen Li %F pmlr-v238-gong24b %I PMLR %P 4276--4284 %U https://proceedings.mlr.press/v238/gong24b.html %V 238 %X Distribution matching can be used to learn invariant representations with applications in fairness and robustness. Most prior works resort to adversarial matching methods but the resulting minimax problems are unstable and challenging to optimize. Non-adversarial likelihood-based approaches either require model invertibility, impose constraints on the latent prior, or lack a generic framework for distribution matching. To overcome these limitations, we propose a non-adversarial VAE-based matching method that can be applied to any model pipeline. We develop a set of alignment upper bounds for distribution matching (including a noisy bound) that have VAE-like objectives but with a different perspective. We carefully compare our method to prior VAE-based matching approaches both theoretically and empirically. Finally, we demonstrate that our novel matching losses can replace adversarial losses in standard invariant representation learning pipelines without modifying the original architectures—thereby significantly broadening the applicability of non-adversarial matching methods.
APA
Gong, Z., Usman, B., Zhao, H. & I Inouye, D.. (2024). Towards Practical Non-Adversarial Distribution Matching . Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:4276-4284 Available from https://proceedings.mlr.press/v238/gong24b.html.

Related Material