Adversarial Feature Augmentation for Unsupervised Domain Adaptation

Abstract

Recent works showed that Generative Adversarial Networks (GANs) can be successfully applied in unsupervised domain adaptation, where, given a labeled source dataset and an unlabeled target dataset, the goal is to train powerful classifiers for the target samples. In particular, it was shown that a GAN objective function can be used to learn target features indistinguishable from the source ones. In this work, we extend this framework by (i) forcing the learned feature extractor to be domain-invariant, and (ii) training it through data augmentation in the feature space, namely performing feature augmentation. While data augmentation in the image space is a well established technique in deep learning, feature augmentation has not yet received the same level of attention. We accomplish it by means of a feature generator trained by playing the GAN minimax game against source features. Results show that both enforcing domain-invariance and performing feature augmentation lead to superior or comparable performance to state-of-the-art results in several unsupervised domain adaptation benchmarks.

Transferrable Prototypical Networks for Unsupervised Domain Adaptation

“…Recently, [30] combines adversarial learning with discriminative feature learning for unsupervised domain adaptation. Most recently, [32] extends domain discriminator by learning domain-invariant feature extractor and performing feature augmentation.…”

Cluster Alignment With a Teacher for Unsupervised Domain Adaptation

“…Using a teacher model for labeling data is inspired by the impressive consistency-based methods in semi-supervised learning (SSL) [16,42]. Recent attempts to apply SSL techniques in UDA include [6,41,46]. CAT differs from these previous works in that CAT exploits the discriminative class-conditional structures in both the alignment and classification procedures while they focus on improving the classifier for the target domain by implementing the cluster assumption [3].…”

Improving protein function prediction with synthetic feature samples created by generative adversarial networks

“…n_estimators: [10,20,30,40,50] max_depth: [1,2,3,None] max_features: [1, "auto", "log2", None], criterion: ["gini", "entropy"]…”

“…The data augmentation task is also an area where GANs show a great potential to achieve good performance. Most of the existing work on GAN-based data augmentation methods also focus on the image processing tasks, such as image classification [28] [29] [30]. For example, Frid-Adar, et.…”