Discriminative Feature Alignment: Improving Transferability of Unsupervised Domain Adaptation by Gaussian-guided Latent Alignment

23 Jun 2020  ·  Jing Wang, Jiahong Chen, Jianzhe Lin, Leonid Sigal, Clarence W. de Silva ·

In this study, we focus on the unsupervised domain adaptation problem where an approximate inference model is to be learned from a labeled data domain and expected to generalize well to an unlabeled data domain. The success of unsupervised domain adaptation largely relies on the cross-domain feature alignment... Previous work has attempted to directly align latent features by the classifier-induced discrepancies. Nevertheless, a common feature space cannot always be learned via this direct feature alignment especially when a large domain gap exists. To solve this problem, we introduce a Gaussian-guided latent alignment approach to align the latent feature distributions of the two domains under the guidance of the prior distribution. In such an indirect way, the distributions over the samples from the two domains will be constructed on a common feature space, i.e., the space of the prior, which promotes better feature alignment. To effectively align the target latent distribution with this prior distribution, we also propose a novel unpaired L1-distance by taking advantage of the formulation of the encoder-decoder. The extensive evaluations on nine benchmark datasets validate the superior knowledge transferability through outperforming state-of-the-art methods and the versatility of the proposed method by improving the existing work significantly. read more

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Domain Adaptation ImageCLEF-DA DFA-ENT Accuracy 89.1 # 4
Domain Adaptation ImageCLEF-DA DFA-SAFN Accuracy 90.2 # 2
Domain Adaptation MNIST-to-USPS DFA-MCD Accuracy 98.6 # 1
Domain Adaptation MNIST-to-USPS DFA-ENT Accuracy 97.9 # 5
Transfer Learning Office-Home DFA-ENT Accuracy 69.2 # 1
Transfer Learning Office-Home DFA-SAFN Accuracy 69.1 # 2
Domain Adaptation SVHN-to-MNIST DFA-MCD Accuracy 98.9 # 2
Domain Adaptation SVHN-to-MNIST DFA-ENT Accuracy 98.2 # 4
Domain Adaptation SYNSIG-to-GTSRB DFA-MCD Accuracy 97.5 # 1
Domain Adaptation SYNSIG-to-GTSRB DFA-ENT Accuracy 96.8 # 2
Domain Adaptation USPS-to-MNIST DFA-MCD Accuracy 96.6 # 7
Domain Adaptation USPS-to-MNIST DFA-ENT Accuracy 96.2 # 9

Methods


No methods listed for this paper. Add relevant methods here