Deep Learning Gets an Outstanding Treatment

A Comprehensive and Accessible Introduction to Deep Learning

“Python Deep Learning, 3rd edition, by Ivan Vasilev” is a solid book on deep learning that has reached a high level of maturity in its third edition, unlike many other books. This book has many unique features.

The book starts with a very good introduction to machine learning and its main models in supervised learning (linear regression, logistic regression, SVM, decision trees, random forests, gradient boosting) and unsupervised learning (clustering, K-means). His explanation of SVM is particularly insightful.

He then covers all the different areas of training machine models with a very good explanation of overfitted models. A simple neural network with one layer is then presented with an introduction to PyTorch, where an example is shown.

The next chapter, the second chapter, gives both a mathematical and conceptual understanding of deep learning. I have a bachelor’s degree in Physics and I am used to how equations are shown and explained in Physics books. I have always been dismayed by how many Machine Learning and Deep Learning textbooks treat equations, just displaying them without further information about them, which actually shows a weakness in the book author’s understanding of the underlying concepts.

This book is fortunately different. Ivan does a fascinating job with the mathematics, a beautiful subject, unfortunately so unknown to many. He starts with the basics of calculus and linear algebra with just what you need to know, with clear explanations, and moves progressively. The equations are all well explained. He also explains how training NNs and linear/logistic regressions have a lot in common. This is followed by the next chapter with a full conceptual explanation of deep learning with an example in PyTorch.

Computer vision is covered next in two chapters. The first one deals with convolutional networks and the second one with advanced computer vision applications (transfer learning, object detection, image segmentation). The explanations are clear, authoritative, and insightful. Examples are given not only in PyTorch, but also in TensorFlow by using Keras.

Natural Language Processing (NLP) and Transformers are then covered in the next four chapters. Starting from NLP and RNN to attention mechanism and transformers to large language models and their advanced applications. Because of the excellent mathematics background given in chapter 2 and the conceptual explanations presented in chapter 3, the material is easy to understand, although the coverage is (fortunately) rather comprehensive. You will have a truly solid understanding of the material.

Chapter 10 is the last one, where the important subject of MLOps is well treated with a beautiful coverage and clarifying examples.

As you can see, this is a book that will give you a deep, solid understanding of deep learning models: the conceptual and mathematical basis of deep learning models with their applications in computer vision and NLP and their deployment in production. I highly recommend it. Ivan Vasilev has done a truly great job.

You can buy this book at Amazon.com at:
https://amzn.to/3Rw38Ug