Short-Term Memory (LSTM) is a cell-like recurrent neural network architecture that
was designed by data scientist to model temporal sequences and their long-range dependencies more accurately than conventional Recurrent Neural Network .
Long Short-Term Memory (LSTM) is used in a sequence of data or in other words a data that is characterized with change in time. An example of such data is Image Captioning, Video Frame, Music and so on.
The LSTM cell keep track of time state. The cell consists of gates which are:
- Read Gate
- Write Gate
- Keep Gate