Supervised Contrastive Learning (paper review)

Review of paper by Prannay Khosla, Piotr Teterwak, Chen Wang et al (Google Research), 2020

Originally published in Deep Learning Reviews on May 12, 2020.

The authors used contrastive loss, which has recently been shown to be very effective at learning deep neural network representations in the self-supervised setting, for supervised learning, and achieved better results than those obtained with cross-entropy loss for ResNet-50 and ResNet-200.

What can we learn from this paper?