In this book, optimization is portrayed as a procedure. It is not realistic to draw out a thorough theoretical model and utilize traditional algorithmic theory and/or mathematical optimization in many practical applications because the environment is too complex.

This book offers a solid machine-learning approach that combines elements of mathematical optimization, game theory, and learning theory: an optimization approach that gains knowledge as more aspects of the problem are observed. This process-based approach to optimization has produced some truly amazing results in modeling and systems that are now a part of our everyday lives.

This widely used graduate-level textbook is based on the author's "Theoretical Machine Learning" course at Princeton University.