Certainty and Uncertainty Guided Active Domain Adaptation

TL;DR

This paper introduces a collaborative active domain adaptation framework that leverages both uncertain and confident target samples, significantly improving adaptation performance by reducing the search space and effectively utilizing confident predictions.

Contribution

It proposes a novel framework combining Gaussian Process-based Active Sampling and Pseudo-Label-based Certain Sampling to enhance active domain adaptation.

Findings

Outperforms state-of-the-art ADA methods on Office-Home and DomainNet datasets.

Incorporating confident predictions reduces the search space and improves adaptation.

The method effectively balances uncertain and confident samples for better domain alignment.

Abstract

Active Domain Adaptation (ADA) adapts models to target domains by selectively labeling a few target samples. Existing ADA methods prioritize uncertain samples but overlook confident ones, which often match ground-truth. We find that incorporating confident predictions into the labeled set before active sampling reduces the search space and improves adaptation. To address this, we propose a collaborative framework that labels uncertain samples while treating highly confident predictions as ground truth. Our method combines Gaussian Process-based Active Sampling (GPAS) for identifying uncertain samples and Pseudo-Label-based Certain Sampling (PLCS) for confident ones, progressively enhancing adaptation. PLCS refines the search space, and GPAS reduces the domain gap, boosting the proportion of confident samples. Extensive experiments on Office-Home and DomainNet show that our approach outperforms state-of-the-art ADA methods.