Recurrent neural network

A recurrent neural network (RNN) is a special type of artificial neural network (a computer program) that can remember things it has seen before. RNNs do this by connecting information from the past (called a hidden state) with new information. Because of this, RNNs are good at processing data in a sequence, like understanding language, translating between languages, or predicting what might come next in a sequence.

References

↑ Zhang, Aston; Lipton, Zachary C.; Li, Mu; Smola, Alexander J. (2023). "9.4. Recurrent Neural Networks". Dive into Deep Learning. Cambridge University Press. ISBN 9781009389433.
↑ Lipton, Zachary C.; Berkowitz, John; Elkan, Charles (17 October 2015). "Sequence to sequence learning with neural networks". arXiv:1506.00019v4 [cs.LG].

This short article about science can be made longer. You can help Wikipedia by adding to it.

[1] Zhang, Aston; Lipton, Zachary C.; Li, Mu; Smola, Alexander J. (2023). "9.4. Recurrent Neural Networks". Dive into Deep Learning. Cambridge University Press. ISBN 9781009389433.

[2] Lipton, Zachary C.; Berkowitz, John; Elkan, Charles (17 October 2015). "Sequence to sequence learning with neural networks". arXiv:1506.00019v4 [cs.LG].