Formal education is great because it gives you the opportunity to bounce ideas off of your professor and peers, put what you learn into practice in controlled environments, and earn a credential to display on your resume. While sometimes costly, it adds much needed discipline and guidance to an educational journey. That said, another great way to learn more about AI & machine learning — and one that’s less expensive, lower-tech, and that you can start today — is reading.
Here, we’ve assembled a list of the best machine learning books, whether you’re just looking to break into the field, already in the field and looking to progress, want to learn more about how to implement machine learning solutions at your company, or want to gain more context about the economic, social, and political implications of machine learning.
All prices were current on Amazon at the time of writing.
Artificial Intelligence: A Modern Approach, 4th Edition. By Peter Norvig and Stuart Russell. Pearson; 1152 pages; $170.66 on Amazon
We begin with the standard textbook for anyone looking to break into artificial intelligence or machine learning, written by Google’s former director of research and search quality (Norvig) and a world-renowned Berkeley computer science professor (Russell). Within, you’ll find up-to-date overviews of some of the most important topics in the field, including deep learning, transfer learning, reinforcement learning, and natural language processing. While it might not get you coding your own machine learning algorithms, if you want an authoritative account of the theory behind artificial intelligence and machine learning practice, this is your book. And if you don’t want to take our word for it, trust the hundreds of professors who assign it to their elementary machine learning and artificial intelligence students.
Introduction to Machine Learning with Python: A Guide for Data Scientists. By Andreas C. Müller and Sarah Guido. O’Reilly Media; 398 pages; $53.62 on Amazon
The first of several O’Reilly entries on the list is Andreas Müller and Sara Guido’s introduction, but don’t be misled by the title: although this book is great for those looking to get smart on machine learning to work in data science, it’s also a valuable resource for future machine learning engineers and other ML professionals. Emphasizing the practical machine learning tools needed to succeed in industry, Müller and Guido offer step-by-step guidance that will help even the complete beginner install essential Python libraries and tools and start writing machine learning algorithms. What the authors don’t include, however, is much explanation of the math behind these algorithms, instead recommending readers turn to resources like Trevor Hastie, Robert Tibshirani, and Jerome Friedman’s The Elements of Statistical Learning (see below).
Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. By Aurélien Géron. O’Reilly Media; 850 pages; $67.10 on Amazon
Géron’s book, also published by O’Reilly, offers a deeper dive into machine learning than Müller and Guido’s introduction while still remaining accessible to those with some skill in Python. Focusing on the Scikit-Learn, Keras, and TensorFlow software libraries, Géron guides his reader through what end-to-end machine learning projects should look like before covering machine learning topics as classification, linear regression, decision trees, dimensionality reduction, and support vector machines. For absolute beginners, we wouldn’t recommend this as your first stop, but if you already have some introductory exposure, or have had the opportunity to learn Python or another programming language, this book is a great way to level-up your machine learning skills.
The Elements of Statistical Learning: Data Mining, Inference, and Prediction. By Trevor Hastie, Robert Tibshirani, and Jerome Friedman. Springer; 767 pages; $73.57 on Amazon
Less a machine learning book than a statistics text, The Elements of Statistical Learning nevertheless offers a crucial overview of the mathematics at the heart of unsupervised and supervised learning, with an emphasis on data mining. Cover-to-cover readers can expect to come away with a better understanding of the statistical concepts working behind topics like neural networks, different methods of regression, and Bayesian reasoning — more likely, however, that the book serves as a quick but detailed reference to tackle statistical questions that come up as you work through other resources.
Becoming a Data Head: How to Think, Speak and Understand Data Science, Statistics and Machine Learning. By Alex J. Gutman and Jordan Goldmeier. Wiley; 272 pages; $24.49 on Amazon
Less a machine learning textbook than a crash-course in data literacy, in Becoming a Data Head data scientists Ajex Gutman and Jordan Goldmeier explain what exactly data is, what kinds of questions aspiring machine learning engineers and data scientists should ask to adopt a data-mindset, and how they can sound smart when talking numbers. They also give accessible overviews of topics like supervised learning and unsupervised learning that can help even the data-averse understand machine learning.
Deep Learning with Python. By François Chollet. Manning. 504 pages. $39.49 on Amazon
For those looking to dip their toes into deep learning, look no further than this ground-up introduction written by the creator of the Keras software library. Chollet offers a clear and deliberate path into coding neural networks with Keras, beginning first with a conceptual introduction to machine learning and deep learning, then introducing the relevant mathematics and software, and finally offering step-by-step guides to a variety of deep learning techniques and applications, including recurrent neural networks (RNNs) for natural language processing, covnets for computer vision, and generative adversarial networks (GANs). Always focused on practical application, Chollet closes the body of the text with best practices for the real world.
An alternative to Chollet’s book that focuses less on coding with Python and more on basic concepts, Deep Learning by Ian Goodfellow, Yoshua Bengio, and Aaron Courville kicks off with relevant applied mathematics for deep learning (linear algebra, probability and information theory), then covers through deep learning basics (neural network architecture, training, optimization) and applications (computer vision, speech recognition, natural language processing), and finally dives into the latest topics in deep learning research (autoencoders, representation learning, deep generative models).
Pattern Recognition and Machine Learning. By Christopher M. Bishop. Springer; 738 pages; $109.99 on Amazon
In this book, Microsoft lab director and computer science professor Christopher Bishop uses probability theory, decision theory, and information theory to introduce machine learning applications for pattern recognition. Techniques and concepts covered by Bishop include reinforcement learning, graphical models such as the decision tree, and Bayesian neural networks. Readers with only a basic knowledge of multivariable calculus and linear algebra will be able to make their way through Bishop’s dense and technical writing, but it’s best suited for those with more significant applied mathematics and computer science background.
Age of Invisible Machines: A Practical Guide to Creating a Hyper Automated Ecosystem of Intelligence Digital Workers. By Robb Wilson. Wiley. 288 pages; $24.33 on Amazon
This entry in our list is not so much for the aspiring data scientist or machine learning engineer as it is for the women and men of corporate management. In Age of Invisible Machines — a Wall Street Journal best seller — technologist Robb Wilson offers practical ways that business executives can put machine learning solutions into practice to reinvigorate their companies and put themselves on a path towards hyperautomation. Wilson centers his book around conversational AI, a natural language processing technology reliant on machine learning, sharing learnings from his experience at the helm of OneReach.ai on how business leaders can implement conversational AI solutions to increase efficiency, cut expenses, and provide better value to users.
HBR’s 10 Must Reads on AI, Analytics, and the New Machine Age. By Harvard Business Review. 192 pages; $17.49 on Amazon
Another entry for corporate players, HBR’s collection of the best of its coverage of AI, data science, and machine learning provides key insights into how businesses are leveraging technologies like blockchain, autonomous vehicles, and augmented reality to maximize profits and other business goals, disrupt existing markets, and provide new value to customers worldwide. Notable contributors include Thomas H. Davenport, professor of IT and management at Babson College; Paul R. Daugherty, chief technology officer at Accenture; Chris Anderson, head of TED; and Marco Iansiti and Karin R. Lakhani, professors at Harvard Business School.
Economic, Social, and Political
Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. By Cathy O’Neil. Crown; 272 pages; $14.39 on Amazon
Drawing case studies from public school systems, political campaigns, baseball, and the job market, data science blogger Cathy O’Neil illustrates what she calls “weapons of math destruction,” or WMDs: big data models encoded with harmful bias, prejudice, or ignorance. Equal parts enjoyable and alarming, O’Nei’s book is a crucial (and early) intervention in the blind faith sometimes accorded big data and data mining.
The Alignment Problem: Machine Learning and Human Values. By Brian Christian. W. W. Norton & Company; 496 pages; $28.95 on Amazon
In The Alignment Problem, programmer and writer Brian Christian similarly tackles the fallibility of the machine learning algorithm — its tendency to perpetuate bias, inequity, and other types of social harm — by examining the issue through thematic lenses like “transparency,” “fairness,” and “inference.” Included in Christian’s exploration of algorithmic unfairness is a glimpse into the steps already being taken by so-called “first responders” to rectify this unfairness and realign machine learning with human values.
The Equality Machine: Harnessing Digital Technology for a Brighter, More Inclusive Future. By Orly Lobel. PublicAffairs; 368 pages; $29.49 on Amazon
For a more optimistic view on the future of machine learning, look to tech policy scholar Orly Lobel’s The Equality Machine, an Economist Best Book of 2022. In it, Lobel dispels the notion that machine learning and other digital technologies are fated to merely increase inequality and other social and environmental ills, instead exploring the ways that machine learning and other digital technologies are in some cases best suited to diagnose and address them. With diverse examples from eBay’s disparity detection algorithm to “marriage markets,” Lobel argues powerfully for approaching algorithmic responsibility questions with nuance: not sacrificing progress in pursuit of perfection.
Prediction Machines: The Simple Economics of Artificial Intelligence. By Ajay Agrawal, Joshua Gans, and Avi Goldfarb. Harvard Business Review Press; 272 pages; $30.00 on Amazon
In Prediction Machines, three economists working at the University of Toronto’s Rotman School of Management demystify the rise of machine learning in business by recasting it as a more efficient and cost-effective mode of predictive data analytics. Complementing their incisive analyses with actionable steps for businesses, the authors cut through the noise to demonstrate how machine learning can remove uncertainty and increase productivity.
Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. By Kate Crawford. Yale University Press; 336 pages; $20.99 on Amazon
Microsoft researcher and USC Annenberg professor Kate Crawford’s Atlas of AI reckons with the environmental and political impacts of machine learning, uncovering for readers the hidden material and labor capital needed to power the algorithms written by machine learning engineers and data scientists. This is an especially good read for those considering entering machine learning or data science who want to better understand their work in a global context.
For those looking to enter or level up in the field of machine learning, there’s never been better resources. From primers, to advanced tutorials, to analyses of machine learning’s economic, political, and social implications, the machine learning publishing world is loaded with books to help you boost your understanding and move forward.