Robust Contrastive Learning Using Negative Samples with Diminished Semantics

Unsupervised learning has recently made exceptional progress because of the development of more effective contrastive learning methods.