Unsupervised Image Classification for Deep Representation Learning
This work addresses the problem of scaling unsupervised learning to large datasets for researchers and practitioners, though it appears incremental as it builds on existing deep clustering and contrastive learning methods.
The paper tackles the scalability issue of deep clustering in unsupervised visual representation learning by proposing a simpler framework that avoids embedding clustering, achieving comparable performance on ImageNet and demonstrating generalization to downstream tasks like object detection and semantic segmentation.
Deep clustering against self-supervised learning is a very important and promising direction for unsupervised visual representation learning since it requires little domain knowledge to design pretext tasks. However, the key component, embedding clustering, limits its extension to the extremely large-scale dataset due to its prerequisite to save the global latent embedding of the entire dataset. In this work, we aim to make this framework more simple and elegant without performance decline. We propose an unsupervised image classification framework without using embedding clustering, which is very similar to standard supervised training manner. For detailed interpretation, we further analyze its relation with deep clustering and contrastive learning. Extensive experiments on ImageNet dataset have been conducted to prove the effectiveness of our method. Furthermore, the experiments on transfer learning benchmarks have verified its generalization to other downstream tasks, including multi-label image classification, object detection, semantic segmentation and few-shot image classification.