Patterns, Predictions, And Actions: A story about machine learning

Livro digital

Título:
Patterns, Predictions, And Actions: A story about machine learning

Autor:
Moritz Hardt, Benjamin Recht

Categoria:
Tecnologia > IA

Doador:
Raffaello D. N.

Sinopse:
Machine learning is often sold as a shortcut from data to decision, but the real challenge is deciding what to predict, what to optimize, and when a prediction should become an action. Its table of contents makes that progression explicit: from Introduction and Fundamentals of prediction, through Supervised learning, Representations and features, Optimization, and Generalization, and then onward to Deep learning, Datasets, Causality, and Sequential decision making. The book is authored by Moritz Hardt and Benjamin Recht, and it moves like a guided tour through the core ideas that shape modern machine learning. Along the way it does more than cover standard learning algorithms. It connects empirical risk minimization, gradient methods, stability, benchmark design, and causal inference, while also showing where prediction breaks down when real-world decisions depend on interventions, confounding, and feedback. The result is a broad, carefully structured introduction to machine learning as both a technical and scientific discipline. Readers get not only the standard toolbox, but also a clearer sense of why datasets age, why generalization is subtle, and why actions can require a different mindset than predictions. It is a strong choice for anyone who wants a conceptual map of the field, not just isolated techniques.