Efficient and Robust Pseudo-Labeling for Unsupervised Domain Adaptation

Abstract

Unsupervised domain adaptation is to transfer knowledge from an annotated source domain to a fully-unlabeled target domain. The conventional methods consider the data which exceed a certain threshold of confidence as pseudo-labeled data for the target domain, and thus choosing the appropriate threshold affects the target performance. In this paper, we propose a new confidence-based weighting scheme for obtaining pseudo-labels and an adaptive threshold adjustment strategy to provide sufficient and accurate pseudo-labels throughout the training. To be precise, our confidence-based weighting scheme generates pseudo-labels to have a different contribution based on the confidence, which makes the performance less sensitive to the threshold. Also, the proposed adaptive threshold adjustment method chooses the threshold according to the degree of adaptation of a network to the target domain, and thus obviates the need for an exhaustive search for the appropriate threshold. Experimental results on a digit classification task show that the proposed methods efficiently utilizes the pseudo-labels to preserve sufficiency and accuracy.