언어정보처리를 위한 데이터베이스 의미론 체계의 구축

Abstract

Lee, Kiyong, Jungha Hong, and Chongmin Lee. 2002. Constructing a Database Semantics System for Linguistic Information Processing. Korean Journal of Linguistics, 27-3, 417-438. The purpose of this paper is to show with computational implementation how a database-theoretic semantic system named DBSem can be constructed in line with model-theoretic computational semantics for processing linguistic information. DBSem consists of an ordinary relational database management system like PostgreSQL with a database, DB, and a user-interface LIPS for processing linguistic information. For the present paper, LIPS is the core of the system. It analyzes input linguistic expressions through a syntactic module like KoSyn for processing Korean sentences, representing resultant semantic information like argument structures, parameter binding, and temporal specifications in attribute-value matrix formats. LIPS then converts these semantic representations into linked tabular formats and stores them all in DB. Consisting of background and other types of relevant knowledge, DB functions as a model or frame for interpreting input information and also making ordinary inferences possible. For testing purposes, some sample Korean sentences will be analyzed with their resultant semantic representations. The resulting analyses will then be stored through some auxiliary program into DB and, on its basis, some queries will be carried out in PostgreSQL.