This is a recurrent neural network, aka the LSTM Neural Network. The output of the previous step is fed into the current step in RNN. Hochreiter and Schmidhuber created the LSTM Neural Network. It addressed the issue of RNN long-term dependencies, in which the RNN cannot predict words stored in long-term memory but can make more accurate predictions based on recent information. RNN does not provide efficient performance as the gap length increases. By default, LSTM can keep the information for a long time. It is used for time-series data processing, prediction, and classification.