Deep Learning in dealing with time series data

Institution: University of Ottawa
Category: Faculty of Engineering
Language: English

Course Description

A recurrent neural network (RNN) is a class of artificial neural networks where connections between nodes can create a cycle, allowing output from some nodes to affect subsequent input to the same nodes. This makes them applicable to tasks of time series data, such as connected handwriting recognition, speech recognition or text-to-speech synthesis. RNNs come in many variants, including Modern Hopfield Neural Network (HNN), Restricted Hopfield Networks, Long Short-Term Memory (LSTM), or Gated Recurrent Unit (GRU). In this lesson, two examples are given, one is using the RHN as the auto-associative memory and the other is using the LSTM to predict the emission of nitrous oxide.