Self-supervised learning

A technique to use modified “selfs” to train a Machine learning system. For instance, one can ask a ML system to discriminate image pairs that are extracted from the same image or those that are coming from different images.

Application to Graph embedding: Bielak2021graph

HaoChen2021provable

Tamkin2021DABS: DABS: A Domain-Agnostic Benchmark for Self-Supervised Learning