BOOK · [4391]
Learning Deep Architectures for AI
Technology
A monograph examining the motivations and principles behind learning algorithms for deep architectures in artificial intelligence. It argues that representing high-level abstractions in vision, language, and other AI tasks requires deep architectures composed of multiple layers of non-linear operations. The book analyzes algorithms such as Deep Belief Networks and related unsupervised methods that emerged after 2006, explaining their success, identifying challenges, and outlining open problems and future research directions.