scientific article; zbMATH DE number 1070884

From MaRDI portal
Publication:4357201





This book is an introductory textbook intended for those that want to obtain a broad insight in the different mathematical approaches currently in use in Artificial Intelligence. The book covers predicate logic, resolution, including prolog, nonmonotonic and default reasoning, semantic nets, Bayesian networks, fuzzy and believe logics, neural networks, decision trees and search strategies. The book spends a lot of pages on placing a certain approach in context, giving motivating examples and showing weakenesses. It also stresses that human thinking and AI are quite remote. Also bibliographical sketches of the major researchers gives the book an almost entertaining value. Necessarily, the book can only present the basics of each approach. In predicate logic, the notion `completeness' is mentioned, but it is not proven. For resolution binary resolution is introduced, but there is no reference to all those refined forms of resolution. This makes the book very suitable for a quick tour along different approaches, but it has the dangerous aspect of suggesting that Artificial Intelligence is rather superficial and does not require or generate deep mathematics. As all subjects touched upon in this book have -- and certainly will have -- a wide body of results behind then, such a broad book appears a little odd, in a similar way an introductory textbook on mathematics, touching upon mathematical analysis, operations research, statistics and computability theory is quite uncommon.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4357201)