Re-energizing Domain Discriminator with Sample Relabeling for Adversarial Domain Adaptation

TL;DR

This paper introduces RADA, a novel method that dynamically relabels target samples to strengthen the domain discriminator, significantly improving feature alignment in unsupervised domain adaptation.

Contribution

We propose a dynamic relabeling strategy to re-energize the domain discriminator, enhancing adversarial training effectiveness in UDA.

Findings

RADA outperforms existing methods on multiple benchmarks.

Dynamic relabeling improves domain classifier strength.

Enhanced feature alignment leads to better target domain performance.

Abstract

Many unsupervised domain adaptation (UDA) methods exploit domain adversarial training to align the features to reduce domain gap, where a feature extractor is trained to fool a domain discriminator in order to have aligned feature distributions. The discrimination capability of the domain classifier w.r.t the increasingly aligned feature distributions deteriorates as training goes on, thus cannot effectively further drive the training of feature extractor. In this work, we propose an efficient optimization strategy named Re-enforceable Adversarial Domain Adaptation (RADA) which aims to re-energize the domain discriminator during the training by using dynamic domain labels. Particularly, we relabel the well aligned target domain samples as source domain samples on the fly. Such relabeling makes the less separable distributions more separable, and thus leads to a more powerful domain classifier w.r.t. the new data distributions, which in turn further drives feature alignment. Extensive experiments on multiple UDA benchmarks demonstrate the effectiveness and superiority of our RADA.