Introducing Recurrent Neural Networks

This chapter will introduce you to the theoretical side of the recurrent neural network (RNN) model. Gaining knowledge about what lies behind this powerful architecture will give you a head start on mastering the practical examples that are provided later in the book. Since you may often find yourself in a situation where a critical decision for your application is needed, it is essential to be aware of the building parts of this model. This will help you act appropriately for the situation.

The prerequisite knowledge for this chapter includes basic linear algebra (matrix operations). A basic knowledge in deep learning and neural networks is also a plus. If you are new to that field, I would recommend first watching the great series of videos made by Andrew Ng (https://www.youtube.com/playlist?list=PLkDaE6sCZn6Ec-XTbcX1uRg2_u4xOEky0); they will help you make your first steps so you are prepared to expand your knowledge. After reading the chapter, you will be able to answer questions such as the following:

• What is an RNN?
• Why is an RNN better than other solutions?
• How do you train an RNN?
• What are some problems with the RNN model?