Self-Supervised Learning

Self-Supervised Learning or SSL is a type of Unsupervised Learning where manually labeled data is not required. Instead, data is modified in an automated way that creates labels from which a model can learn—for example, masking random words in a sequence and predicting the missing words. It's based on the idea that unlabelled data is easier to come by than labelled and resembles how infants learn by observing adults around them.