The Hundred-Page Machine Learning Book
About this book
Burkov compresses the field into a book you can read in a weekend, with the mathematics present but held to what is genuinely needed. It covers supervised learning from linear and logistic regression through decision trees, support vector machines and nearest neighbours, then neural networks, then the practical matters of feature engineering, regularisation, hyperparameter tuning and evaluation metrics. Later chapters touch unsupervised learning, dimensionality reduction and applied topics such as recommender systems and ranking. The density is deliberate, so it works as a map of the territory rather than as a tutorial. Read it to hold a credible technical conversation with the people who do build models.
Description via Product Digest.
Topics
Related concepts
- Scaling and emergenceWhy capability improves predictably with size while individual abilities appear abruptly, and what that does to the job of testing a model.
- How models are trained to behavePre training, supervised fine tuning, reinforcement learning from human feedback and constitutional methods, and why a helpful model is a manufactured artefact that can be unmade the same way.
- InterpretabilityWhat the field has established about the inside of a model, what it has not, and why a system that explains itself is not explaining itself.
- AI, machine learning and deep learningArtificial intelligence, machine learning and deep learning are three nested categories, and the difference between them decides what a team can promise about a system.
- Foundation models and how they workTokens, embeddings and transformers explained for somebody who has to make decisions about a model, with nothing in it about building one.