Contrastive Self-Supervised Learning is a method in representation learning that aims at training a neural network to learn how to discriminate different semantic features in images. In this post, we explore the latest research in contrastive self-supervised made around the year 2022.