Machine learning is considered by many to be the future of statistics and computer engineering as it reshapes customer service, design, banking, medicine, manufacturing, and hosts of other disciplines and industries. It is hard to overstate its impact on the world so far and the changes it will bring about in the coming years and decades. Of the multitude of machine learning methods applied by professionals, such as penalized regression, random forests, and boosted trees, perhaps the most excitement-inducing is deep learning.
Deep learning has revolutionized computer vision and natural language processing, and researchers are still finding new areas to transform with the power of neural networks. Its most profound impact is often seen in efforts to replicate the human experience, such as the aforementioned vision and language processing, and also audio synthesis and translations. The math and concepts underlying deep learning can seem daunting, unnecessarily deterring people from getting started.
The authors of Deep Learning Illustrated challenge the traditionally perceived barriers and impart their knowledge with ease and levity, resulting in a book that is enjoyable to read. Much like the other books in this series—R for Everyone, Pandas for Everyone, Programming Skills for Data Science, and Machine Learning with Python for Everyone—this book is welcoming and accessible to a broad audience from myriad backgrounds. Mathematical notation is kept to a minimum and, when needed, the equations are presented alongside understandable prose. The majority of insights are augmented with visuals, illustrations, and Keras code, which is also available as easy-to-follow Jupyter notebooks.
Jon Krohn has spent many years teaching deep learning, including a particularly memorable presentation at the New York Open Statistical Programming Meetup—the same community from which he launched his Deep Learning Study Group. His mastery of the subject shines through in his writing, giving readers ample education while at the same time inviting them to be excited about the material. He is joined by Grant Beyleveld and Aglaé Bassens who add their expertise in applying deep learning algorithms and skillful drawings.
Deep Learning Illustrated combines theory, math where needed, code, and visualizations for a comprehensive treatment of deep learning. It covers the full breadth of the subject, including densely connected networks, convolutional neural nets, recurrent neural nets, generative adversarial networks, and reinforcement learning, and their applications. This makes the book the ideal choice for someone who wants to learn about neural networks with practical guidance for implementing them. Anyone can, and should, benefit from, as well as enjoy, their time spent reading along with Jon, Grant, and Aglaé.