Introduction to Deep Learning

Charniak's guide teaches deep learning through hands-on coding. Each step asks readers to build a working program, so the core ideas come from writing and running code. The applications covered include computer vision, natural language processing, and reinforcement learning.

The book surveys the main architectures and methods of the field: feed-forward and convolutional networks, word embeddings, recurrent networks, sequence-to-sequence models, deep reinforcement learning, and unsupervised approaches. The examples are written in TensorFlow, a free, open-source machine learning library. An early chapter introduces TensorFlow and how it works with Python, a popular programming language.

The author has spent a long career in artificial intelligence research, with a focus on natural language processing. He believes that computer science is best learned by writing programs, and the book follows that belief. Every chapter contains a programming project, a set of exercises, and suggestions for further reading.

Readers should be comfortable with linear algebra, multivariable calculus, probability, and statistics, and should know basic Python. The text suits both undergraduate and graduate courses, and working practitioners can use it as a reference.