Statistical Language Learning

Eugene Charniak breaks new ground in artificial intelligenceresearch by presentingstatistical language processing from an artificial intelligence point of view in a text forresearchers and scientists with a traditional computer science background.New, exacting empiricalmethods are needed to break the deadlock in such areas of artificial intelligence as robotics,knowledge representation, machine learning, machine translation, and natural language processing(NLP). It is time, Charniak observes, to switch paradigms. This text introduces statistical languageprocessing techniques ;word tagging, parsing with probabilistic context free grammars, grammarinduction, syntactic disambiguation, semantic wordclasses, word-sense disambiguation ;along with theunderlying mathematics and chapter exercises.Charniak points out that as a method of attacking NLPproblems, the statistical approach has several advantages. It is grounded in real text and thereforepromises to produce usable results, and it offers an obvious way to approach learning: "one simplygathers statistics."Language, Speech, and Communication