ICML 2019oral769 citations
On Learning Invariant Representations for Domain Adaptation
Han Zhao, Remi Tachet Des Combes, Kun Zhang, Geoffrey Gordon
Abstract
Due to the ability of deep neural nets to learn rich representations, recent advances in unsupervised domain adaptation have focused on learning domain-invariant features that achieve a small error on the source domain. The hope is that the learnt representation, together with the hypothesis learnt from the source domain, can generalize to the target domain. In this paper, we first construct a simple counterexample showing that, contrary to common belief, the above conditions are not sufficient to guarantee successful domain adaptation. In particular, the counterexample exhibits
BibTeX
@InProceedings{pmlr-v97-zhao19a,
title = {On Learning Invariant Representations for Domain Adaptation},
author = {Zhao, Han and Combes, Remi Tachet Des and Zhang, Kun and Gordon, Geoffrey},
booktitle = {Proceedings of the 36th International Conference on Machine Learning},
pages = {7523--7532},
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/zhao19a/zhao19a.pdf},
url = {https://proceedings.mlr.press/v97/zhao19a.html},
abstract = {Due to the ability of deep neural nets to learn rich representations, recent advances in unsupervised domain adaptation have focused on learning domain-invariant features that achieve a small error on the source domain. The hope is that the learnt representation, together with the hypothesis learnt from the source domain, can generalize to the target domain. In this paper, we first construct a simple counterexample showing that, contrary to common belief, the above conditions are not sufficient to guarantee successful domain adaptation. In particular, the counterexample exhibits