A recurrent neural network (RNN) is a deep neural network trained on sequential or time series data to create a machine learning model that can make sequential predictions or conclusions based on sequential inputs. RNNs can be used for language translation, NLP, sentiment analysis, speech recognition, and image captioning