Self-training for Few-shot Transfer Across Extreme Task Differences

All few-shot learning techniques must be pre-trained on a large, labeled "base dataset". In problem domains where such large labeled datasets are not available for pre-training (e.g., X-ray images), one must resort to pre-training in a different "source" problem domain (e.g., ImageNet), which can be very different from the desired target task. Traditional few-shot and transfer learning techniques fail in the presence of such extreme differences between the source and target tasks. In this paper, we present a simple and effective solution to tackle this extreme domain gap: self-training a source domain representation on unlabeled data from the target domain. We show that this improves one-shot performance on the target domain by 2.9 points on average on a challenging benchmark with multiple domains. Despite progress in visual recognition, training recognition systems for new classes in novel domains requires thousands of labeled training images per class and several hours of compute. For example, to train a recognition system for different kinds of pneumonia in chest X-rays, one would have to get radiologists to label thousands of X-ray images, and then spend several hours to train a neural network on high-end GPUs.