Unsupervised embedding learning aims to extract good representation from data without the need for any manual labels, which has been a critical challenge in many supervised learning tasks. This paper proposes a new unsupervised embedding approach, called Super-AND, which extends the current state-of-the-art model. Super-AND has its unique set of losses that can gather similar samples nearby within a low-density space while keeping invariant features intact against data augmentation. Super-AND outperforms all existing approaches and achieves an accuracy of 89.2% on the image classification task for CIFAR-10. We discuss the practical implications of this method in assisting semi-supervised tasks.
@article{arxiv.2002.12158,
title = {A Comprehensive Approach to Unsupervised Embedding Learning based on AND Algorithm},
author = {Sungwon Han and Yizhan Xu and Sungwon Park and Meeyoung Cha and Cheng-Te Li},
journal= {arXiv preprint arXiv:2002.12158},
year = {2020}
}