Modified Long Short-Term Memory Recurrent Neural Network Architectures

Long Short-Term Memory (LSTM) is a specific recurrent neural network (RNN) architecture that was designed to model temporal sequences and their long-range dependencies more accurately than conventional RNNs. In this paper, we explore LSTM RNN architectures and made some changes for its better performance. LSTM RNNs are more effective than DNNs. Here, we have changed the gates calculation and also have removed some unnecessary features of standard LSTM architecture. This architecture makes more effective use of model parameters than the others considered, converges quickly, and outperforms a deep feed forward neural network having an order of magnitude more parameters.