EXPLUS - A SEMANTIC PARSING SYSTEM FOR JAPANESE SENTENCES.

被引:0
|
作者
Tanaka, Hozumi
Sato, Taisuke
Motoyoshi, Fumio
机构
关键词
Computer programming languages - Pattern recognition systems;
D O I
暂无
中图分类号
学科分类号
摘要
EXPLUS is a system to extract semantic structure from Japanese sentences. The extraction process is divided into two phases: 1) Generation of a parsed tree from an input sentence using both Japanese syntactic rules and a parsing system called Extended LINGOL; the tree is considered only as in intermediate form of semantic structure; 2) Derivation of semantic structure from the parsed tree. The major concern is to develop a technique to extract an identical semantic structure from sentences whose meanings are almost the same. The notion of UNIT-to-UNIT interaction is introduced to analyze dependency relations between words in a sentence. In the interaction, production rules are activated to guide the analysis. SRL is proposed as acknowledge representation language that incorporates some useful features in KRL developed by D. G. Bobrow and T. Winograd. Features of SRL are discussed.
引用
收藏
页码:236 / 240
相关论文
共 50 条