Contrastive Adaptation Network for Unsupervised Domain Adaptation

About

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.

Guoliang Kang, Lu Jiang, Yi Yang, Alexander G Hauptmann · 2019

Related benchmarks

Task	Dataset	Result
Image Classification	Office-31	Average Accuracy90.6	357
Unsupervised Domain Adaptation	Office-Home (test)	Average Accuracy68.9	347
Domain Adaptation	Office-31	Average Accuracy90.6	187
Domain Adaptation	Office-31 unsupervised adaptation standard	Accuracy (A to W)94.5	162
Object Classification	VisDA synthetic-to-real 2017	Mean Accuracy87.2	139
Unsupervised Domain Adaptation	Office-31	A->W Accuracy94.5	116
Unsupervised Domain Adaptation	ImageCLEF-DA	Average Accuracy89.6	114
Unsupervised Domain Adaptation	VisDA unsupervised domain adaptation 2017	Mean Accuracy87.2	103
Image Classification	Office-31 (test)	Avg Accuracy90.6	93
Image Classification	VisDA-C (test)	Mean Accuracy87.2	76