Design and Implementation of a Lexical Disambiguation System

摘要

The paper reports on a research project which covers the development of an English-Chinese machine translator using an Artificial Intelligence approach. For the machine translation part, the Augmented Transition Network (ATN) approach is used. In this way however only the syntax of the language is used for translation of the sentences. Errors can occur based on semantic ambiguities, lexical ambiguities or structural ambiguities of the human language. Therefore, the paper attempts to find another solution for correcting errors which occur after syntax based translation using semantic approach of the language. The Chinese and the English sentence will pass through a Montague-inspired semantic interpreter which results in two logic translations mapping the text of the natural language to a representation of its meaning. Hereafter these logic sentences will be compared and eventual errors will be fed back to the ATN based translator, after which errors can be corrected.