Contrastive Adaptation Network for Unsupervised Domain Adaptation

TL;DR

This paper introduces the Contrastive Adaptation Network (CAN), a novel approach for unsupervised domain adaptation that explicitly models intra-class and inter-class discrepancies to improve feature discrimination and domain alignment.

Contribution

The paper proposes a new contrastive metric and an end-to-end training strategy for UDA, enhancing class-aware domain adaptation performance.

Findings

CAN outperforms state-of-the-art methods on Office-31 and VisDA-2017.

CAN produces more discriminative features for target domain data.

Experimental results validate the effectiveness of the proposed approach.

Abstract

Unsupervised Domain Adaptation (UDA) makes predictions for the target domain data while manual annotations are only available in the source domain. Previous methods minimize the domain discrepancy neglecting the class information, which may lead to misalignment and poor generalization performance. To address this issue, this paper proposes Contrastive Adaptation Network (CAN) optimizing a new metric which explicitly models the intra-class domain discrepancy and the inter-class domain discrepancy. We design an alternating update strategy for training CAN in an end-to-end manner. Experiments on two real-world benchmarks Office-31 and VisDA-2017 demonstrate that CAN performs favorably against the state-of-the-art methods and produces more discriminative features.