Slowly, painfully, inexorably, AI gets more and more useful. AI techniques are now showing up in everything from recommendation systems to news agents, sensors to speech recognition -- and, increasingly, in mainstream business and technical software, too. This book takes AI out of the traditional “AI ghetto,” introducing its algorithms and techniques to “typical” programmers who can now derive great value from them.
M. Tim Jones covers pretty much every useful category of AI programming: neural networks, expert systems, fuzzy logic, genetic algorithms, rules-based systems, and more. While a few of the sample applications are theoretical, most are very practical, and drawn straight from the real world.
For example, Jones introduces Adaptive Resonance Theory (ART) in the context of e-commerce recommender systems (“you liked ‘Book A,’ you’ll probably like ‘Book Q,’ too”). There’s a chapter on the recently developed “ant algorithms,” which attempt to mimic the techniques blind ants use to navigate complex environments and get back home safely. (Exceptionally useful for solving highly dynamic problems, such as routing in rapidly changing networks.)
Jones shows how neural networks can be used in creating intelligent game behaviors; how intelligent agents are being used in today’s content and news filtering systems; how artificial life works; and how hidden Markov models can be used to fabricate text that resembles existing literary works. (Maybe they’ll never generate new Shakespeare, but Tom Clancy? Definitely possible.)
All the source code’s on CD-ROM. It’s tested to run under Red Hat Linux (6.1 and up) or any recent version of Windows running the free Cygwin Unix emulator. Bill Camarda
Bill Camarda is a consultant, writer, and web/multimedia content developer. His 15 books include Special Edition Using Word 2000 and Upgrading & Fixing Networks for Dummies, Second Edition.