The neural network architectures such as multi-layers perceptron (MLP) were trained using the current inputs only. We did not consider previous inputs when generating the current output. In other words, our systems did not have … Continue reading Recurrent Neural Networks — Part 1