Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation

Xiang Jiang, Qicheng Lao, Stan Matwin, Mohammad Havaei
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:4816-4827, 2020.

Abstract

We present an approach for unsupervised domain adaptation{—}with a strong focus on practical considerations of within-domain class imbalance and between-domain class distribution shift{—}from a class-conditioned domain alignment perspective. Current methods for class-conditioned domain alignment aim to explicitly minimize a loss function based on pseudo-label estimations of the target domain. However, these methods suffer from pseudo-label bias in the form of error accumulation. We propose a method that removes the need for explicit optimization of model parameters from pseudo-labels. Instead, we present a sampling-based implicit alignment approach, where the sample selection is implicitly guided by the pseudo-labels. Theoretical analysis reveals the existence of a domain-discriminator shortcut in misaligned classes, which is addressed by the proposed approach to facilitate domain-adversarial learning. Empirical results and ablation studies confirm the effectiveness of the proposed approach, especially in the presence of within-domain class imbalance and between-domain class distribution shift.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-jiang20d, title = {Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation}, author = {Jiang, Xiang and Lao, Qicheng and Matwin, Stan and Havaei, Mohammad}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {4816--4827}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/jiang20d/jiang20d.pdf}, url = {https://proceedings.mlr.press/v119/jiang20d.html}, abstract = {We present an approach for unsupervised domain adaptation{—}with a strong focus on practical considerations of within-domain class imbalance and between-domain class distribution shift{—}from a class-conditioned domain alignment perspective. Current methods for class-conditioned domain alignment aim to explicitly minimize a loss function based on pseudo-label estimations of the target domain. However, these methods suffer from pseudo-label bias in the form of error accumulation. We propose a method that removes the need for explicit optimization of model parameters from pseudo-labels. Instead, we present a sampling-based implicit alignment approach, where the sample selection is implicitly guided by the pseudo-labels. Theoretical analysis reveals the existence of a domain-discriminator shortcut in misaligned classes, which is addressed by the proposed approach to facilitate domain-adversarial learning. Empirical results and ablation studies confirm the effectiveness of the proposed approach, especially in the presence of within-domain class imbalance and between-domain class distribution shift.} }
Endnote
%0 Conference Paper %T Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation %A Xiang Jiang %A Qicheng Lao %A Stan Matwin %A Mohammad Havaei %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-jiang20d %I PMLR %P 4816--4827 %U https://proceedings.mlr.press/v119/jiang20d.html %V 119 %X We present an approach for unsupervised domain adaptation{—}with a strong focus on practical considerations of within-domain class imbalance and between-domain class distribution shift{—}from a class-conditioned domain alignment perspective. Current methods for class-conditioned domain alignment aim to explicitly minimize a loss function based on pseudo-label estimations of the target domain. However, these methods suffer from pseudo-label bias in the form of error accumulation. We propose a method that removes the need for explicit optimization of model parameters from pseudo-labels. Instead, we present a sampling-based implicit alignment approach, where the sample selection is implicitly guided by the pseudo-labels. Theoretical analysis reveals the existence of a domain-discriminator shortcut in misaligned classes, which is addressed by the proposed approach to facilitate domain-adversarial learning. Empirical results and ablation studies confirm the effectiveness of the proposed approach, especially in the presence of within-domain class imbalance and between-domain class distribution shift.
APA
Jiang, X., Lao, Q., Matwin, S. & Havaei, M.. (2020). Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:4816-4827 Available from https://proceedings.mlr.press/v119/jiang20d.html.

Related Material