BERT is an abbreviation for Bidirectional Encoder Representations from Transformers, which is a transformer-based vectorising and embedding model. It was created by Google researchers. Because of the use of transformers and an attention mechanism, it has a remarkably high accuracy.
