The Norm Must Go On: Dynamic Unsupervised Domain Adaptation by Normalization
This addresses the challenge of dynamic adaptation in scenarios like autonomous driving with scarce data, though it is incremental as it builds on existing batch normalization methods.
The paper tackles the problem of continuous adaptation to distribution shifts in unsupervised domain adaptation by proposing Dynamic Unsupervised Adaptation (DUA), which adapts batch normalization statistics with minimal unlabeled data, achieving competitive results with less than 1% of target domain data and low computational overhead.
Domain adaptation is crucial to adapt a learned model to new scenarios, such as domain shifts or changing data distributions. Current approaches usually require a large amount of labeled or unlabeled data from the shifted domain. This can be a hurdle in fields which require continuous dynamic adaptation or suffer from scarcity of data, e.g. autonomous driving in challenging weather conditions. To address this problem of continuous adaptation to distribution shifts, we propose Dynamic Unsupervised Adaptation (DUA). By continuously adapting the statistics of the batch normalization layers we modify the feature representations of the model. We show that by sequentially adapting a model with only a fraction of unlabeled data, a strong performance gain can be achieved. With even less than 1% of unlabeled data from the target domain, DUA already achieves competitive results to strong baselines. In addition, the computational overhead is minimal in contrast to previous approaches. Our approach is simple, yet effective and can be applied to any architecture which uses batch normalization as one of its components. We show the utility of DUA by evaluating it on a variety of domain adaptation datasets and tasks including object recognition, digit recognition and object detection.