BYOL -- Bootstrap Your Own Latent

It is a method of machine learning where the model learns from the supervisory signal of the data unlike supervised learning where separate labels are specified for each observation. It is also known as Representation Learning. Note, the model's learned representation is used for downstream tasks like BERT, where language models are used for downstream tasks like text classification. Here, we can use Linear classifiers along with a learned self-supervised model for prediction. Recently, self-supervised learning has seen a great surge in the number of papers getting published for few obvious reasons including the availability of unlabelled data.