Recurrent Neural Networks
A recurrent neural network (RNN) is a type of deep learning model built for sequential data. It processes inputs one step at a time and carries a hidden state forward, so each step is interpreted in light of what came before. This makes RNNs useful for time series and natural language.
At each step, an RNN combines the current input with its memory of previous steps and updates that memory. Plain RNNs struggle to retain information over long sequences, so variants such as long short-term memory (LSTM) and gated recurrent units (GRU) add gates that control what to keep and what to forget.
In manufacturing, an RNN can read a stream of sensor readings from a motor and predict when vibration or temperature patterns suggest an upcoming failure.
RNNs differ from transformers, which process an entire sequence in parallel using attention. Transformers have largely replaced RNNs for language tasks because they train faster on large datasets, though RNNs remain in use for some time-series work.