Release Of BERT By Google | Vibepedia

In October 2018, Google researchers introduced BERT, a language model that dramatically improved the state of the art for large language models, and has since become a ubiquitous baseline in natural language processing experiments. BERT's innovative approach to learning contextual, latent representations of tokens has found applications in many NLP tasks. The release of BERT has spawned the study of 'BERTology', which attempts to interpret what is learned by BERT, and has been influenced by earlier models such as [[elmo|ELMo]] and [[gpt-2|GPT-2]].