The Master Algorithm — The book Bill Gates suggests you read to better understand AI | by Bernardo Pereira | Nov, 2020


The book has 10 chapters, every with the aim of exposing an concept or exploring a selected matter. I made a decision to summarize every chapter in order that you can get an concept of how the book is structured and the matters it covers.

1. The Machine Learning Revolution: Machine Learning is launched by way of examples of some functions, the guarantees of what it may well obtain, and current obstacles. At the top of the chapter, the writer lays out the questions this book tries to reply.

2. The Master Algorithm: Here, the books’ central speculation is offered:

‘All knowledge — past, present, and future — can be deduced from data by a single, universal learning algorithm.’

This common studying algorithm is what Pedro Domingos refers to as The Master Algorithm. Pedro Domingos refers to the work that has been carried out by 5 totally different colleges of thought or what he calls the ‘5 tribes of machine learning’: Symbolists, Connectionists, Evolutionaries, Bayesians, and Analogizers.

This lays the bottom for the next 5 chapters, every diving into one of many tribes.

3. Hume’s downside of induction: Dive into the Symbolists who incorporate pre-existing information within the studying course of. Their grasp algorithm is induction.

4. How does your mind work: Dive into the Connectionists, to whom studying requires understanding how the mind works. More particularly, how neuron’s connections adapt primarily based on new proof. Their grasp algorithm is backpropagation.

5. Evolution: natures studying algorithm: Dive into the Evolutionaries who’re impressed by mom nature and pure choice. Their grasp algorithm is genetic programming.

6. In the church of Reverend Bayes: Dive into the Bayesians. This tribe is generally involved about uncertainty as all information has some extent of uncertainty to it. Their grasp algorithm is the Bayes Theorem.

7. You are what you purpose: Dive into the Analogizers, who consider that the important thing to studying is to acknowledge similarities between occasions and use these to infer additional similarities. Their grasp algorithm is the Support Vector Machine.

8. Learning with no instructor: Introduction to unsupervised studying, a method the writer believes to be vital for the Master Algorithm. This part contains transient explanations of algorithms equivalent to k-means, Principal Component Analysis, and Chunking. Moreover, it skims by way of reinforcement studying.

9. The items of the puzzle fall in place: The writer proposes methods of mixing the totally different algorithms and methodologies so as to create the Master Algorithm. Some of the present tasks that intention to achieve this are offered.

10. This is the world on Machine Learning: The writer shares his imaginative and prescient of the that means and real-world affect of what has been mentioned. He does so by devising episodes of our on a regular basis lives and the way they’re going to be impacted.


Source hyperlink

Write a comment