Black-box Source-free Domain Adaptation via Two-stage Knowledge Distillation
This addresses privacy concerns in domain adaptation by eliminating the need for source data access, though it is incremental as it builds on existing knowledge distillation techniques.
The paper tackles the problem of black-box source-free domain adaptation, where only source model outputs and target data are available, by proposing a two-stage knowledge distillation method that trains a target model from scratch and uses a new student model to reduce error accumulation, achieving strong results on three cross-domain segmentation tasks.
Source-free domain adaptation aims to adapt deep neural networks using only pre-trained source models and target data. However, accessing the source model still has a potential concern about leaking the source data, which reveals the patient's privacy. In this paper, we study the challenging but practical problem: black-box source-free domain adaptation where only the outputs of the source model and target data are available. We propose a simple but effective two-stage knowledge distillation method. In Stage \uppercase\expandafter{\romannumeral1}, we train the target model from scratch with soft pseudo-labels generated by the source model in a knowledge distillation manner. In Stage \uppercase\expandafter{\romannumeral2}, we initialize another model as the new student model to avoid the error accumulation caused by noisy pseudo-labels. We feed the images with weak augmentation to the teacher model to guide the learning of the student model. Our method is simple and flexible, and achieves surprising results on three cross-domain segmentation tasks.