Semi-supervised learning made simple with self-supervised clustering
This work addresses the challenge of leveraging partial labels in real-world scenarios, offering an incremental improvement to existing semi-supervised learning methods.
The paper tackled the problem of semi-supervised learning by proposing a simple method to adapt clustering-based self-supervised models like SwAV or DINO into semi-supervised learners, achieving state-of-the-art performance on CIFAR100 and ImageNet.
Self-supervised learning models have been shown to learn rich visual representations without requiring human annotations. However, in many real-world scenarios, labels are partially available, motivating a recent line of work on semi-supervised methods inspired by self-supervised principles. In this paper, we propose a conceptually simple yet empirically powerful approach to turn clustering-based self-supervised methods such as SwAV or DINO into semi-supervised learners. More precisely, we introduce a multi-task framework merging a supervised objective using ground-truth labels and a self-supervised objective relying on clustering assignments with a single cross-entropy loss. This approach may be interpreted as imposing the cluster centroids to be class prototypes. Despite its simplicity, we provide empirical evidence that our approach is highly effective and achieves state-of-the-art performance on CIFAR100 and ImageNet.