Class-tested and up-to-date textbook for introductory courses on information retrieval.

Class-tested and up-to-date textbook for introductory courses on information retrieval.

This volume is concerned with how ambiguity and ambiguity resolution are learned, that is, with the acquisition of the different representations of ambiguous linguistic forms and the knowledge necessary for selecting among them in context. Schütze concentrates on how the acquisition of ambiguity is possible in principle and demonstrates that particular types of algorithms and learning architectures (such as unsupervised clustering and neural networks) can succeed at the task. Three types of lexical ambiguity are treated: ambiguity in syntactic categorisation, semantic categorisation, and verbal subcategorisation. The volume presents three different models of ambiguity acquisition: Tag Space, Word Space, and Subcat Learner, and addresses the importance of ambiguity in linguistic representation and its relevance for linguistic innateness.

Abstract: "We present a new approach to disambiguating syntactically ambiguous words in context, based on Variable Memory Markov (VMM) models. In contrast to fixed-length Markov models, which predict based on fixed-length histories, variable memory Markov models dynamically adapt their history length based on the training data, and hence may use fewer parameters. In a test of a VMM based tagger on the Brown corpus, 95.81% of tokens are correctly classified."