Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation
This work addresses domain adaptation for computer vision tasks, offering a novel method to enhance feature discriminability, though it appears incremental relative to existing adversarial approaches.
The paper tackles the problem of suboptimal performance in unsupervised domain adaptation by proposing Drop to Adapt (DTA), which uses adversarial dropout to learn discriminative features, achieving consistent improvements in image classification and semantic segmentation tasks.
Recent works on domain adaptation exploit adversarial training to obtain domain-invariant feature representations from the joint learning of feature extractor and domain discriminator networks. However, domain adversarial methods render suboptimal performances since they attempt to match the distributions among the domains without considering the task at hand. We propose Drop to Adapt (DTA), which leverages adversarial dropout to learn strongly discriminative features by enforcing the cluster assumption. Accordingly, we design objective functions to support robust domain adaptation. We demonstrate efficacy of the proposed method on various experiments and achieve consistent improvements in both image classification and semantic segmentation tasks. Our source code is available at https://github.com/postBG/DTA.pytorch.