Introduction

Abstract

In this chapter, we provide a whirlwind introduction of the history of deep neural networks (also known as deep learning), positioned in a broader scope of machine learning and artificial intelligence. We then focus on a specific research direction of deep neural networks—incorporating structural information of data into the design of network architectures. This motivates the key contribution of the book, a tree-based convolutional neural network (TBCNN), that performs the convolution operation over tree structures. Finally, we provide an overview of this book.