AI Glossary

BERT

Also called Bidirectional Encoder Representations from Transformers

What does “BERT” mean?

BERT: Encoder-only transformer released by Google in 2018 that reads text in both directions at once to understand each word in context. It is not built to write long text, but it became a standard base for classification, search ranking and other language-understanding tasks.

BERT visual guide: Encoder-only transformer released by Google in 2018 that reads text in both directions at once to understand each word in context. It is not built to write long text, but it became a standard base for classification, search ranking and other language-understanding tasks.
Visual guide Open full size (new tab)

What is an example of “BERT”?

Google said in 2019 that it had started using BERT to better understand the meaning of search queries.

Keep learning

See also

EncoderTermTransformerTermEmbedding modelTermSelf-supervised learningTerm

Same topic

Convolutional neural networkTermAutoregressive modelTermDecoderTermAttention mechanismTermDense modelTerm