Machine learning is one of the most popular disciplines in the field of Artificial Intelligence. Whether you’re a student, seasoned developer, or titan of industry—reading books on machine learning and artificial intelligence will ensure you stay current on one of the world’s fastest-developing fields.
Machine learning has proven useful in distilling great actionable insights from large modern datasets. This computing revolution has seen machine learning propelled to the forefront of academia and commercial research programs.
Finding the best machine learning and Artificial Intelligence books is a daunting task. Between academic papers, textbooks, practical guides, and narrative texts there are thousands of options. This article reflects the effort to narrow the list and compile a consensus offering of the best titles available for learning machine learning.
The titles listed here were selected from a list of recommended titles across more than a dozen “best machine learning” oriented articles. The sources of these articles were chosen based on ML & AI-focused keyword searches in Google with relevancy to book recommendations. The listings here represent the titles most frequently recommended from all of these URLs. For more details on the selection criteria for this list read this project analysis.
The Best Machine Learning Books
The titles here, representing an aggregate of recommendations, reflect a diverse presentation of information. Some are textbooks, others code-inclusive practical guides, and others narrative discussions of such topics as ethics and philosophy of machine learning and artificial intelligence. They are presented in order of descending frequency of recommendation—the first being the most frequently recommended title.
The Hundred-Page Machine Learning Book
Author(s): Andriy Burkov
The Hundred-page machine learning book delivers what the title implies—a simple, condensed, no-nonsense survey of machine learning. This book introduces an industry-wide survey of concepts, algorithms, and applications of machine learning. It offers colorized charts, graphs, and illustrations to depict in-use ML technologies.
This book should not be regarded as rigorous academic or technical instruction and focuses on width rather than depth of study—through some existing machine learning knowledge is highly recommended before reading.
Author(s): Ian Goodfellow, Yoshua Bengio, Aaron Courville
This book provides a rich detailing of deep learning theory by practitioners in the field. It has a technical orientation but refrains from diving too deeply into the advanced discussion until later chapters. This book is intended for a more advanced audience interested in the foundations of Deep Learning.
Critics of this book convey their impression that it was written without enough personality and reads a bit tersely. Advocates assert the elegance in which a survey of the field is presented and note the light nature in which advanced theories are introduced.
Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems
Author(s): Aurélien Géron
This book presents readers with a gentle introduction to machine learning concepts, starting low and going slowly into more advanced applications and algorithms. The authors provide practical examples using Python, with heavy use on Keras (via TensorFlow) and Scikit-Learn.
Critics note that there have been some printing issues such that code examples and figures don’t appear in the physical book copy. Advocates often cite the accessibility, practicality, and completeness in which this book is written. For anyone familiar with the quality of O’Reilly texts—this one measures up to the expected quality and lands firmly among the most popular machine learning books.
Superintelligence: Paths, Dangers, Strategies
Author(s): Nick Bostrom
Superintelligence is a narrative survey of potential pitfalls of a projected future where Artificial Intelligence develops a robust general intelligence surpassing that of humans. This book offers footnotes, bibliographic references, and industry insights related to a possible future where AI has brought misfortune to humankind. The authors discuss potential negative outcomes of AI systems with respect to economics, ethics, and possible development routes to such a disaster.
Critics note this book presents information in a slightly repetitive manner and in a format much too technical for the average reader. Advocates note the level of detail reinforces the authors’ arguments, and that the warnings expressed are much needed as we enter an age with deeper AI integration in our collective daily lives.
Machine Learning for Absolute Beginners
Author(s): Oliver Theobald
This book delivers what one expects after reading the title: a survey of machine learning catered to the technical awareness of an absolute beginner. This title provides a survey of concepts, vocabulary, and applications for common machine learning technologies. This book is not intended for those looking for advanced study or even considerable elaboration on specific topics. Rather, Machine Learning for Absolute Beginners serves more like a roadmap to the ML industry, its developing technology, and the algorithms at its heart.
Critics express the opinions this book contains not enough elaboration, doesn’t adequately express mathematical concepts, or in some cases becomes a bit too technical. Advocates argue this to be one of the best introduction machine learning books for providing a survey of information to guide one’s study. Its worth noting this book is now in the third edition and, as such, one might find a lower number of available reviews given the majority are associated with previous editions.
Artificial Intelligence: A Modern Approach
This is a textbook-format introduction to the theory and practice of many emerging applications of artificial intelligence. It’s designed as a two-semester course suitable for graduate or undergraduate level comprehension of AI subject matter. Topics surveyed include Intelligent Agents, Uncertainty, probability, neural networks, reinforcement learning, and much, much, more.
Advocates of this book note that it does well to cover nearly the entire field of AI in a comprehensive manner, the accessible writing style and prose used by the authors, commonly suggest its inclusion as a “classic” among AI texts. Critics note this book has often been received in a poor quality of physical binding (publisher issue, not content,) that the indexing and table of contents don’t offer any help in accessing the material, and a lack of practical examples (heavy on theory.)
Author(s): Stuart J. Russel, Peter Norvig
This is a textbook format book offering a comprehensive introduction to machine learning concepts and algorithms. It doesn’t strive to be the authority on any specific topic within the machine learning field but rather to provide readers and students with a wide view of the field. The topics and discussions are presented in a concise and accessible format along with many illustrative examples providing a range of accommodation for learning styles.
Advocates of this book note the success of its authors in providing a discussion on a broad range of topics, the quality and utility of the practical examples discussed, and the timelessness of format within a fast-paced field. Critics argue some of the examples are a bit out-of-date, the price-tag being a bit high (it is a textbook), and the omission of certain newer ML concepts such as support vector machines (SVMs).
Author(s): Tom M. Mitchell
Deep Learning with Python
This book on machine learning was authored by Keras creator and Google AI researcher François Chollet. The book presents topics in high-level discussion with accompanying practical examples with a strong emphasis on the Keras library developed for the Python programming language. This book was written for those of any technical level interested in the field of deep learning and focuses more on code, examples, and goes light on mathematical notation and discussion.
Advocates of this book note the quality and accessibility provided in the introduction and the overall approachability of the text. In addition, the practicality of examples, utility of code examples, and the friendly weaving through more advanced topics like generative adversarial networks (GANs) are often cited as a strong selling point. Critics note many of the examples in the book require little problem solving or critical thinking and are more follow the instructions. In addition, many libraries referenced in this text are outdated, unavailable, or greatly restructured (e.g. Keras is now part of TensorFlow.)
Author(s): Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani
Machine Learning For Dummies
A somewhat as-expected read for anyone familiar with the for Dummies series but a gem nonetheless. This title covers a beginner-friendly, but still technically sound, introduction to machine learning and artificial intelligence as emerging fields of modern relevancy. It touches on mathematical concepts underpinning machine learning, the applications for common ML algorithms, and even walks the reader through practical steps on implementing machine learning.
Advocates of this title note the accessibility, range of discussion, and relevancy to modern ML and AI technologies in both R and Python. Critics note that the book might not be quite as accessible as the “Dummies” in the title implies presenting more advanced topics than one might expect. Expect a broad survey in technical depths suited for a range of skillsets but also parts that might seem to drag such as installing Python and/or R.
Author(s): John Paul Mueller
Pattern Recognition and Machine Learning
This title presents technical implementations of Bayesian-oriented pattern recognition algorithms in a textbook format. It contains deep discussions of mathematical concepts and is catered towards seasoned developers, those with a reasonably solid mathematical background, and those comfortable with basic concepts of Bayesian statistics.
Advocates of this title note that concepts are introduced in a fairly linear fashion with adequate discussion as the book progresses through examples and problem sets. In addition, the title’s targeting of graduate and Ph.D. level students is well-received by those seeking advanced information. Critics note some mathematical errors are present in the book and that, at times, the discussion could be a bit more concise as new topics are introduced.
Author(s): Christopher M. Bishop
The Singularity Is Near
This title serves as both an education and warning of the potential dangers when humankind reaches the point where AI surpasses human intelligence. The author cites many literary and scientific references to support his points, concepts, and warnings. The premise of the book is the scenario that might result from the compounding growth of technology in the Artificial Intelligence field.
Advocates of this title note how well the author utilizes scientific support for the ideas and concepts presented. The author’s inventiveness, creativity, and ability to segue into creative discussion amidst more daunting technical forecasting. Critics note the concepts presented are hardly new and that the basic premise has been discussed for decades. Additionally, those critical of this title commonly report their surprise at the technical nature of the subject matter whereas a more fictional, Sci-Fi thriller might have been expected.
Author(s): Judea Pearl, Dana Mackenzie
The Master Algorithm
The Master Algorithm is written to be accessible by a general audience and provides a survey of the Machine Learning field starting from its past and working into modern and yet-to-be-realized implementations. The author progresses through many common algorithms, detailing their role in solving complex problems.
The author presents the concept of a “master algorithm” that could solve any problem with enough data. Through the book, the esoteric concept of this idea becomes more feasible as the author weaves their way through technically fluent discussions of the ML field, its future, and current applications. Advocates praise the comprehensive-yet-accessible format while critics note some of the praise for the book’s accessibility may be a bit overzealous as some parts require deeper technical knowledge to grasp.
Author(s): Pedro Domingos
Human + Machine
Human + Machine is a detailed account of how artificial intelligence has become integral to businesses and economic applications on a global scale. Throughout the book. the authors discuss the role AI is currently playing, as well as the role it may play, in companies around the world. This title reads as though the target audience is business leaders and those standing to benefit from deeper insight into the role AI can play in the modern economic terrain.
Advocates of this title note that not only does this book paint a favorable picture for AI’s future role in business but it also provides actionable steps leaders can take today. Additionally, advocates praise the author’s deep experience within the field of business such that their vision of AI is well-substantiated and the discussion practical. Critics of the book note a lack of technical depth related to specific AI technology as well as the author’s choice to establish new terminology to discuss hypotheticals.
Author(s): Paul R. Daugherty, H. James Wilson
Life 3.0: Being Human in the Age of Artificial Intelligence
Life 3.0 is a title written to be accessible to the general public. It seeks to clear the record of AI’s role in the future—one the author feels has been influenced too much by Hollywood. Rather than focusing on any single application or issues surrounding AI, Life 3.0 covers a wide berth of scenarios in which future AI will play a role—from job markets to modern military conflicts, to political and socio-economic applications.
Advocates of this title note the power behind the author’s explanatory style, their use of the Life 1.0, Life 2.0, and Life 3.0 system to paint a picture of our collective progression as a species. Critics of the title note the author seems cemented in their stance of portraying AI as positive and keeping an unbalanced tone of positivity. Additionally, those critical of this book describe the author’s use of fictitious scenarios and lack of adding technical depth weaken the perspectives and arguments posed.
Author(s): Tegmark Max Tegmark Max
Among the search for the best machine learning books several titles came up in searches but either didn’t make it onto recommendations lists or didn’t make it onto enough lists to be included. These books should be regarded as biased recommendations—not derived from the process by which other titles were included:
Machine Learning Yearning
This book is authored by one of the most notable names in the Machine Learning field: Andrew Ng. His online courses are rated among the best ML courses available, his works in the field have helped pioneer some of the most utilized approaches in ML, and his role as co-founder of Google Brain makes him nothing less than an authority on the subject. This title is available for free download from the DeepLearning.ai website in PDF format.
Author: Andrew Ng
About These Choices
The books on this list were chosen by a survey of more than a dozen websites dedicated partially or wholly to the study of machine learning. These websites were identified through a series of Google keyword searches related to machine learning and artificial intelligence interests.
Books mentioned on fewer than three source websites were excluded from this list. The books with the greatest frequency of mention were ranked highest on this list—many reflecting the same frequency.
Readers are urged to read the full analysis of how this article was produced for a full discussion of the criteria for both selection, inclusion, and exclusion as well as the larger process by which the machine learning books on this list were compiled.
Artificial intelligence is a broad field nestled in the heart of computer science. While machine learning is only a subset of AI study it has grown in popularity in recent years. The books on this list are relevant to future CS students seeking to prepare for study, recent graduates looking at doctoral programs, and seasoned professionals sharping their skillsets.
Related: Read this article for a list of general computer science and programming books.
This article has taken to account a consensus of voices from prominent AI and machine learning. This consensus is intended not to serve as the canonical list of the best books out there but, rather, serve as a well-balanced jumping-off point for those seeking more information on machine learning and artificial intelligence. As such, readers are urged to not regard a book’s omission from this list as an indication of a lack of value or any degree of inferiority.