Active Adversarial Domain Adaptation
This addresses domain adaptation for machine learning models when labeled data is scarce in the target domain, but it is incremental as it builds on existing adversarial and sampling techniques.
The paper tackles the problem of domain adaptation when the source domain has many labeled examples but the target domain has few, by proposing an active learning approach that unifies adversarial domain alignment and importance sampling. It shows significant improvements over fine-tuning and other methods, retaining advantages even after hundreds of annotations in tasks like object detection.
We propose an active learning approach for transferring representations across domains. Our approach, active adversarial domain adaptation (AADA), explores a duality between two related problems: adversarial domain alignment and importance sampling for adapting models across domains. The former uses a domain discriminative model to align domains, while the latter utilizes it to weigh samples to account for distribution shifts. Specifically, our importance weight promotes samples with large uncertainty in classification and diversity from labeled examples, thus serves as a sample selection scheme for active learning. We show that these two views can be unified in one framework for domain adaptation and transfer learning when the source domain has many labeled examples while the target domain does not. AADA provides significant improvements over fine-tuning based approaches and other sampling methods when the two domains are closely related. Results on challenging domain adaptation tasks, e.g., object detection, demonstrate that the advantage over baseline approaches is retained even after hundreds of examples being actively annotated.