Continual Source-Free Unsupervised Domain Adaptation
This addresses the challenge of domain adaptation in scenarios where source data is unavailable, which is incremental as it builds on existing SUDA methods.
The paper tackles the problem of catastrophic forgetting in source-free unsupervised domain adaptation (SUDA) by proposing a continual learning framework that adapts to target domains while maintaining source performance without accessing source data, achieving satisfactory results on benchmarks like PACS, Visda-C, and DomainNet.
Existing Source-free Unsupervised Domain Adaptation (SUDA) approaches inherently exhibit catastrophic forgetting. Typically, models trained on a labeled source domain and adapted to unlabeled target data improve performance on the target while dropping performance on the source, which is not available during adaptation. In this study, our goal is to cope with the challenging problem of SUDA in a continual learning setting, i.e., adapting to the target(s) with varying distributional shifts while maintaining performance on the source. The proposed framework consists of two main stages: i) a SUDA model yielding cleaner target labels -- favoring good performance on target, and ii) a novel method for synthesizing class-conditioned source-style images by leveraging only the source model and pseudo-labeled target data as a prior. An extensive pool of experiments on major benchmarks, e.g., PACS, Visda-C, and DomainNet demonstrates that the proposed Continual SUDA (C-SUDA) framework enables preserving satisfactory performance on the source domain without exploiting the source data at all.