Talk to The Master Algorithm like its author wrote you back.
Get the ideas that fit your life — not generic summaries.
- Chat with the book
- Audiobook-style main ideas
- Adapts to your life and goals
- Helps you take action
What it argues
The Master Algorithm is Pedro Domingos's survey of machine learning — the field of computer science that creates algorithms capable of learning from data — organized around a central speculative thesis: that there exists, or may be found, a single master algorithm from which all learning can be derived. Domingos argues that the major schools of machine learning are each working on different facets of the same underlying problem, and that a unified learner that combines their strengths would be capable of deriving any knowledge from data.
The five tribes of machine learning that Domingos describes each embody a different metaphor for learning. Symbolists derive rules from examples using inverse deduction, following in the tradition of logic and formal systems. Connectionists build systems inspired by neural structure, most powerfully expressed in modern deep learning. Evolutionists use simulated evolution — genetic algorithms — to discover solutions by survival of the fittest. Bayesians reason probabilistically under uncertainty, updating beliefs as evidence accumulates. Analogizers generalize from similar examples, expressed most powerfully in support vector machines and k-nearest-neighbor approaches.
What it gets right
- 1.
Machine learning is the programming paradigm that creates algorithms which improve their performance through experience rather than being explicitly programmed to solve specific problems.
- 2.
Five major schools of machine learning — symbolists, connectionists, evolutionists, Bayesians, and analogizers — represent different approaches to the same fundamental problem of learning from data.
- 3.
Deep learning, the most powerful current form of connectionist learning, builds hierarchical representations of data through many layers of artificial neurons, and has achieved human-level or superhuman performance on vision and language tasks.
What it covers
Who wrote it
Pedro Domingos is a professor of computer science at the University of Washington and one of the leading researchers in machine learning. He received his doctorate from UC Irvine and has worked on probabilistic inference, social networks, and the unification of machine learning paradigms. He is the creator of the Markov logic network, which combines probabilistic and logical inference. The Master Algorithm was named one of the best books of 2015 by the Financial Times and Amazon. He is known for his synthetic perspective across the different schools of machine learning.
