The Master Algorithm
How the Quest for the Ultimate Learning Machine Will Remake Our World
Synopsis
Pedro Domingos, a machine learning researcher at the University of Washington, argues that the field's ultimate goal is a single universal learner, a "master algorithm" able to derive any knowledge from data. To get there, he sorts machine learning into five tribes, each with its own master method: symbolists, who use inverse deduction; connectionists, who train neural networks with backpropagation; evolutionaries, who rely on genetic programming; Bayesians, who reason with probabilistic inference; and analogizers, who learn from similarity with methods such as support vector machines. Successive chapters explain each tribe through a familiar idea, from logic and the brain to natural selection, probability and analogy, with few equations.
The later chapters sketch how the five approaches might be unified, including Domingos's own work on Markov logic networks, and consider what pervasive learning machines mean for business, science and personal data, where he proposes data unions to give individuals more control. Written in 2015, the book predates the transformer and the scaling of deep learning that produced today's large language models, so its forecast of a synthesized learner now reads as one hypothesis among several. Some reviewers found its metaphors folksy or confusing. Readers still gain a durable conceptual map of how machines learn and a vocabulary for comparing techniques.
Scorecard
Editorial score
How it scored
Editorial score: 71.1 / 1005 weighted criteria
- 24.0/30
Influence30%
- 21.3/25
Authority25%
- 11.0/20
Practical value20%
- 8.3/15
Currency15%
- 6.5/10
Readability10%
- 71.1/100
Editorial score
How the rank was computed
Rank No. 59 of 500 comes from a composite of 67.9: observable signals 35.9 of 55 points plus editorial review 32.0 of 45 points.
All 11 inputs
| Input | Points |
|---|---|
| Observable signals | 35.9/55 |
| Reach (Amazon ratings volume) | 11.9/16.5 |
| Consensus across best-of lists | 13.8/13.8 |
| Awards and bestseller evidence | 5.0/8.3 |
| Reader rating, sample-adjusted | 4.8/8.3 |
| Amazon sales rank | 0.4/5.5 |
| Recency | 0.0/2.8 |
| Editorial review | 32.0/45 |
| Author authority | 9.6/11.3 |
| Influence | 10.8/13.5 |
| Practical value | 5.0/9.0 |
| Readability | 2.9/4.5 |
| Currency | 3.7/6.8 |
Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.
Verdict
An ambitious, readable map of machine learning's five schools of thought, still useful for intuition though dated on deep learning's rise.
Strengths
- Clear taxonomy of machine learning approaches
- Written by a leading machine learning researcher
- Engaging history of the field's rivalries
Keep in mind
Its 2015 vantage point underplays deep learning's later dominance, and readers without a background in logic or computing may find some chapters dense.
Key takeaways
- Machine learning divides into five tribes (symbolists, connectionists, evolutionaries, Bayesians and analogizers), each with a characteristic algorithm and worldview.
- Every learner combines three components (representation, evaluation and optimization), which gives readers a common way to compare very different techniques.
- Domingos argues that a universal learner combining the five tribes' strengths could, in principle, discover any knowledge derivable from data.
- As learning systems model people more closely, individuals need collective mechanisms, such as data unions, to control how their data is used.
Notable facts
- At the 2016 Code Conference, Bill Gates told the audience to read Pedro Domingos's book. Source (opens in a new tab)
- In 2016 Bill Gates named it, with Nick Bostrom's Superintelligence, one of two books everyone should read to understand AI; in 2018 it was reported on Xi Jinping's bookshelf. Source (opens in a new tab)
About the author
Questions
What is The Master Algorithm about?
Is The Master Algorithm outdated?
Who should read The Master Algorithm?
Related books
- An Introduction to Statistical Learning Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani and Jonathan Taylor
- Reinforcement Learning Richard S. Sutton and Andrew G. Barto
- The Elements of Statistical Learning Trevor Hastie, Robert Tibshirani and Jerome Friedman
- Why Machines Learn Anil Ananthaswamy
- Pattern Recognition and Machine Learning Christopher M. Bishop
- Data Science for Business Foster Provost and Tom Fawcett
Sources
As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.






