Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation

TL;DR

This paper introduces an implicit class-conditioned domain alignment method for unsupervised domain adaptation that effectively handles class imbalance and distribution shift without relying on pseudo-label optimization.

Contribution

It proposes a sampling-based implicit alignment approach that avoids pseudo-label bias and addresses domain-discriminator shortcuts, improving adaptation under challenging class conditions.

Findings

Effective in handling class imbalance and distribution shift

Outperforms existing methods in empirical evaluations

Reduces pseudo-label bias and error accumulation

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 directly. Instead, we present a sampling-based implicit alignment approach, where the sample selection procedure 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 implicit alignment 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.