Integration of Syntactic Analysis and Semantic Interpretation Based on Equivalent Transformation
Hiroshi Mabuchi*, Kiyoshi Akama**, Takahiko Ishikawa***, and Hidekatsu Koike**
*Faculty of Software and Information Science, Iwate Prefectural University, 152-52 Sugo, Takizawa, Iwate 020-0193, Japan
**Research Division of Large-Scale Computational Systems, Information Initiative Center, Hokkaido University, Kita 11, Nishi 5, Kita-ku, Sapporo 060-0811, Japan
***Division of System and Information Engineering, Hokkaido University, Kita 11, Nishi 5, Kita-ku, Sapporo 060-0811, Japan
Making an efficient algorithm for natural language understanding by means of flexible and cooperative interaction between syntactic analysis and semantic interpretation is very difficult. In order to overcome the difficulties, the present paper proposes a new method for designing knowledge processing systems, the computation of which is based on equivalent transformation of declarative descriptions. Basic procedures for syntactic analysis and semantic interpretation are formalized as mutually independent equivalent transformation rules. Rule selection is dynamically determined flexibly during execution by a general principle independent of the domain of sentences.