Domain Agnostic Learning with Disentangled Representations

Xingchao Peng, Zijun Huang, Ximeng Sun, Kate Saenko
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:5102-5112, 2019.

Abstract

Unsupervised model transfer has the potential to greatly improve the generalizability of deep models to novel domains. Yet the current literature assumes that the separation of target data into distinct domains is known a priori. In this paper, we propose the task of Domain-Agnostic Learning (DAL): How to transfer knowledge from a labeled source domain to unlabeled data from arbitrary target domains? To tackle this problem, we devise a novel Deep Adversarial Disentangled Autoencoder (DADA) capable of disentangling domain-specific features from class identity. We demonstrate experimentally that when the target domain labels are unknown, DADA leads to state-of-the-art performance on several image classification datasets.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-peng19b, title = {Domain Agnostic Learning with Disentangled Representations}, author = {Peng, Xingchao and Huang, Zijun and Sun, Ximeng and Saenko, Kate}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {5102--5112}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/peng19b/peng19b.pdf}, url = {https://proceedings.mlr.press/v97/peng19b.html}, abstract = {Unsupervised model transfer has the potential to greatly improve the generalizability of deep models to novel domains. Yet the current literature assumes that the separation of target data into distinct domains is known a priori. In this paper, we propose the task of Domain-Agnostic Learning (DAL): How to transfer knowledge from a labeled source domain to unlabeled data from arbitrary target domains? To tackle this problem, we devise a novel Deep Adversarial Disentangled Autoencoder (DADA) capable of disentangling domain-specific features from class identity. We demonstrate experimentally that when the target domain labels are unknown, DADA leads to state-of-the-art performance on several image classification datasets.} }
Endnote
%0 Conference Paper %T Domain Agnostic Learning with Disentangled Representations %A Xingchao Peng %A Zijun Huang %A Ximeng Sun %A Kate Saenko %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-peng19b %I PMLR %P 5102--5112 %U https://proceedings.mlr.press/v97/peng19b.html %V 97 %X Unsupervised model transfer has the potential to greatly improve the generalizability of deep models to novel domains. Yet the current literature assumes that the separation of target data into distinct domains is known a priori. In this paper, we propose the task of Domain-Agnostic Learning (DAL): How to transfer knowledge from a labeled source domain to unlabeled data from arbitrary target domains? To tackle this problem, we devise a novel Deep Adversarial Disentangled Autoencoder (DADA) capable of disentangling domain-specific features from class identity. We demonstrate experimentally that when the target domain labels are unknown, DADA leads to state-of-the-art performance on several image classification datasets.
APA
Peng, X., Huang, Z., Sun, X. & Saenko, K.. (2019). Domain Agnostic Learning with Disentangled Representations. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:5102-5112 Available from https://proceedings.mlr.press/v97/peng19b.html.

Related Material