ICML 2020poster159 citations

Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation

Xiang Jiang, Qicheng Lao, Stan Matwin, Mohammad Havaei

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.

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.}
}
Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation · ICML 2020