Source-Free Unsupervised Domain Adaptation with Hypothesis Consolidation of Prediction Rationale
This work addresses a challenging domain adaptation problem for machine learning practitioners, offering an incremental improvement by enhancing pseudo-label accuracy through hypothesis consolidation.
This paper tackles the problem of source-free unsupervised domain adaptation, where a model must adapt to a new domain without access to source data or target labels, by proposing a method that consolidates prediction rationales from multiple hypotheses to identify correct pseudo-labels for semi-supervised learning. It achieves state-of-the-art performance in this task, with experimental results showing it can be integrated into existing approaches to improve their performance.
Source-Free Unsupervised Domain Adaptation (SFUDA) is a challenging task where a model needs to be adapted to a new domain without access to target domain labels or source domain data. The primary difficulty in this task is that the model's predictions may be inaccurate, and using these inaccurate predictions for model adaptation can lead to misleading results. To address this issue, this paper proposes a novel approach that considers multiple prediction hypotheses for each sample and investigates the rationale behind each hypothesis. By consolidating these hypothesis rationales, we identify the most likely correct hypotheses, which we then use as a pseudo-labeled set to support a semi-supervised learning procedure for model adaptation. To achieve the optimal performance, we propose a three-step adaptation process: model pre-adaptation, hypothesis consolidation, and semi-supervised learning. Extensive experimental results demonstrate that our approach achieves state-of-the-art performance in the SFUDA task and can be easily integrated into existing approaches to improve their performance. The codes are available at \url{https://github.com/GANPerf/HCPR}.