Self-Supervised Noisy Label Learning for Source-Free Unsupervised Domain Adaptation
This addresses the problem of domain adaptation without access to source data due to privacy and transmission constraints, offering a novel approach that is incremental in combining noisy label learning with self-supervision.
The paper tackles source-free unsupervised domain adaptation, where only a pre-trained model and unlabeled target data are available, by framing it as a noisy label learning problem and proposing a self-supervised method to fine-tune the model. The method achieves state-of-the-art results, surpassing other methods by a large margin, as validated through extensive experiments.
It is a strong prerequisite to access source data freely in many existing unsupervised domain adaptation approaches. However, source data is agnostic in many practical scenarios due to the constraints of expensive data transmission and data privacy protection. Usually, the given source domain pre-trained model is expected to optimize with only unlabeled target data, which is termed as source-free unsupervised domain adaptation. In this paper, we solve this problem from the perspective of noisy label learning, since the given pre-trained model can pre-generate noisy label for unlabeled target data via directly network inference. Under this problem modeling, incorporating self-supervised learning, we propose a novel Self-Supervised Noisy Label Learning method, which can effectively fine-tune the pre-trained model with pre-generated label as well as selfgenerated label on the fly. Extensive experiments had been conducted to validate its effectiveness. Our method can easily achieve state-of-the-art results and surpass other methods by a very large margin. Code will be released.