AI Glossary

Recurrent neural network

Also called RNN

What does “Recurrent neural network” mean?

Recurrent neural network: Recurrent networks process a sequence one step at a time, carrying a hidden state that summarizes what they have seen so far. That makes them natural for text, speech and time series, but they are slow to train in parallel and tend to forget distant information.

Recurrent neural network visual guide: Recurrent networks process a sequence one step at a time, carrying a hidden state that summarizes what they have seen so far. That makes them natural for text, speech and time series, but they are slow to train in parallel and tend to forget distant information.
Visual guide Open full size (new tab)

What is an example of “Recurrent neural network”?

Before transformers, RNNs powered most machine translation and speech recognition systems; their limits with long sequences helped motivate the shift to attention.

Keep learning

See also

Long short-term memoryTermTransformerTermState-space modelTerm

Same topic

Convolutional neural networkTermSelf-attentionTermReasoning modelTermSmall language modelTermPositional encodingTerm