Selective Pseudo-Labeling with Reinforcement Learning for Semi-Supervised Domain Adaptation
This work addresses the challenge of effectively leveraging limited labeled data in the target domain for semi-supervised domain adaptation, which is an incremental improvement for researchers and practitioners working on domain adaptation.
This paper tackles the problem of semi-supervised domain adaptation (SSDA) where existing methods often fail to improve performance despite a few labeled target instances. The authors propose a reinforcement learning-based selective pseudo-labeling method that uses deep Q-learning to select accurate and representative pseudo-labeled instances, achieving superior performance on benchmark datasets.
Recent domain adaptation methods have demonstrated impressive improvement on unsupervised domain adaptation problems. However, in the semi-supervised domain adaptation (SSDA) setting where the target domain has a few labeled instances available, these methods can fail to improve performance. Inspired by the effectiveness of pseudo-labels in domain adaptation, we propose a reinforcement learning based selective pseudo-labeling method for semi-supervised domain adaptation. It is difficult for conventional pseudo-labeling methods to balance the correctness and representativeness of pseudo-labeled data. To address this limitation, we develop a deep Q-learning model to select both accurate and representative pseudo-labeled instances. Moreover, motivated by large margin loss's capacity on learning discriminative features with little data, we further propose a novel target margin loss for our base model training to improve its discriminability. Our proposed method is evaluated on several benchmark datasets for SSDA, and demonstrates superior performance to all the comparison methods.