Augmentation Learning for Semi-Supervised Classification
This addresses the need for adaptable semi-supervised learning in domain-specific image tasks, though it is incremental as it builds upon existing methods like Fixmatch.
The paper tackles the problem of semi-supervised learning methods relying on manually designed data augmentation pipelines that are not transferable to other domains, by proposing a method that automatically selects effective augmentation policies for specific datasets, achieving state-of-the-art results on satellite images and hand-drawn sketches.
Recently, a number of new Semi-Supervised Learning methods have emerged. As the accuracy for ImageNet and similar datasets increased over time, the performance on tasks beyond the classification of natural images is yet to be explored. Most Semi-Supervised Learning methods rely on a carefully manually designed data augmentation pipeline that is not transferable for learning on images of other domains. In this work, we propose a Semi-Supervised Learning method that automatically selects the most effective data augmentation policy for a particular dataset. We build upon the Fixmatch method and extend it with meta-learning of augmentations. The augmentation is learned in additional training before the classification training and makes use of bi-level optimization, to optimize the augmentation policy and maximize accuracy. We evaluate our approach on two domain-specific datasets, containing satellite images and hand-drawn sketches, and obtain state-of-the-art results. We further investigate in an ablation the different parameters relevant for learning augmentation policies and show how policy learning can be used to adapt augmentations to datasets beyond ImageNet.