Adjunction Structure Formalism for English-Korean Machine Translation System

Abstract

In this paper, we present a knowledge representation which is focused on the correspondence of adjunction structure between English and Korean. We acquire the knowledge available in machine translation, from the linguistic structures and data in the two languages. The adjunction structure is the constituent representing the relation between the modifier and the modifier. Our approach is based on the unified method of knowledge description to formalize the correspondence between the two languages. We verify the effectiveness of our method for the purpose of constructing the English-Korean machine translation system.