Source-Free Domain Adaptation via Multi-view Contrastive Learning
This work addresses privacy concerns in domain adaptation by enabling adaptation without access to labeled source data, though it is incremental as it builds on existing SFUDA methods to improve accuracy.
The paper tackles the challenges of low-quality prototype samples and incorrect pseudo-label assignments in Source-Free Unsupervised Domain Adaptation (SFUDA) by proposing a method with a Reliable Sample Memory module, Multi-View Contrastive Learning, and noisy label filtering. It achieves approximately 2% and 6% improvements in classification accuracy over the second-best method and the average of 13 state-of-the-art approaches on benchmark datasets.
Domain adaptation has become a widely adopted approach in machine learning due to the high costs associated with labeling data. It is typically applied when access to a labeled source domain is available. However, in real-world scenarios, privacy concerns often restrict access to sensitive information, such as fingerprints, bank account details, and facial images. A promising solution to this issue is Source-Free Unsupervised Domain Adaptation (SFUDA), which enables domain adaptation without requiring access to labeled target domain data. Recent research demonstrates that SFUDA can effectively address domain discrepancies; however, two key challenges remain: (1) the low quality of prototype samples, and (2) the incorrect assignment of pseudo-labels. To tackle these challenges, we propose a method consisting of three main phases. In the first phase, we introduce a Reliable Sample Memory (RSM) module to improve the quality of prototypes by selecting more representative samples. In the second phase, we employ a Multi-View Contrastive Learning (MVCL) approach to enhance pseudo-label quality by leveraging multiple data augmentations. In the final phase, we apply a noisy label filtering technique to further refine the pseudo-labels. Our experiments on three benchmark datasets - VisDA 2017, Office-Home, and Office-31 - demonstrate that our method achieves approximately 2 percent and 6 percent improvements in classification accuracy over the second-best method and the average of 13 well-known state-of-the-art approaches, respectively.