One distinguishing feature of the 6th edition is the prominent place given to Hidden Markov Models. Indeed, one might have asked for these to have been equally prominent in earlier editions. For several (>10) years, HMMs have been successfully used in various practical applications. Above all, in Automatic Speech Recognition. To often correctly infer the word or phrase that was uttered. The models have made ASRs prominent and for the most part, practical in being used in mass consumer applications. But HMMs in those contexts were not often considered AI per se. Here, the text moves HMM squarely into view, as a valid and vital technique for AI.



Not that the text is restricted to this, of course. It still has a broad introductory coverage of major AI topics. Consider the predicate calculus. Or stochastic methods to infer meaning. [You might consider HMM to be a special type of stochastic method.]



Perhaps the best summary of the book is that it seems attuned to practical applications of AI. The algorithm descriptions and suggested usages aid the porting to contexts where you do not necessarily need the full panoply of AI. The hard AI problems you might leave to others. You can treat this entire text as a good summation of powerful computational algorithms.