Chapter 11
Introducing Recurrent Neural Networks
IN THIS CHAPTER
Understanding the importance of learning data in sequence
Creating image captions and translating languages using deep learning
Discovering the long short-term memory (LSTM) technology
Knowing about possible alternatives to LSTM
This chapter explores how deep learning can deal with information that flows. Reality is not simply changeable, but is changeable in a progressive way that is made predictable by observing the past. If a picture is a static snapshot of a moment in time, a video, consisting of a sequence of related images, is flowing information, and a film can tell you much more than a single photo or a series of photos can. Likewise for short and long textual data (from tweets to entire documents or books) and for all numeric series that represent something occurring along a timeline (for instance, a series the about sales of a product or the quality of the air by day in a city).
This chapter explains a series of new layers, the recurrent networks, and all their improvements, such as the LSTM and GRU layers. ...