Learning Deep Learning: Theory and Practice of Neural Networks, Computer Vision, NLP, and Transformers using TensorFlow

Chapter 9

Predicting Time Sequences with Recurrent Neural Networks

In this chapter, we introduce another important neural network architecture known as the recurrent neural network (RNN). This architecture is useful when doing predictions based on sequential data, and especially for sequences of variable lengths. Before explaining what an RNN is, we provide some context by describing some of the problem types to which RNNs can be applied. We relate these problem types to the tasks we have already encountered in previous chapters.

Up until now, we have applied networks to two main categories of tasks. One was a regression problem in which the network predicted a real-valued variable based on a number of other variables, such as the example of ...