This book is for people with no experience with machine learning and who are looking for an intuition-based, hands-on introduction to deep learning using Python.
Deep Learning for Complete Beginners: A Python-Based Introduction is for complete beginners in machine learning. It introduces fundamental concepts such as classes and labels, building a dataset, and what a model is and does before presenting classic machine learning models, neural networks, and modern convolutional neural networks. Experiments in Python--working with leading open-source toolkits and standard datasets--give the reader hands-on experience with each model and help them build intuition about how to transfer the examples in the book to their own projects.
Readers start with an introduction to the Python language and the NumPy extension that is ubiquitous in machine learning. Prominent toolkits, like sklearn and Keras/TensorFlow are used as the backbone to enable readers to focus on the elements of machine learning without the burden of writing implementations from scratch. An entire chapter on evaluating the performance of models gives the reader the knowledge necessary to understand claims on performance and to know which models are working well and which are not. The book culminates by presenting convolutional neural networks as an introduction to modern deep learning. Understanding how these networks work and how they are affected by parameter choices leaves the reader with the core knowledge necessary to dive into the larger, ever-changing world of deep learning.