Distributed syntactic representations with an application to part-of-speech tagging

The syntactic category of words is represented in a distributed manner by deriving part-of-speech representations from a text corpus by means of a large scale singular value decomposition. The representations are input to an artificial neural network, which is trained on two days of the New York Times News Service, attaining an accuracy of 92%-98% in tagging ambiguous lexical items.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>