JETCAT – Japanese-English Translation Using Corpus-Based Acquisition of Transfer Rules

要旨

Abstract — In this paper we present a rule-based formalism for the acquisition, representation, and application of the transfer knowledge used in a Japanese-English machine translation system. The transfer knowledge is learnt automatically from a parallel corpus by using structural matching between the parse trees of translation pairs. The user can customize the rule base by simply correcting translation results. We have extended the machine translation system with two user-friendly front ends: an MS Word interface and a Web interface. Since our system is mainly intended as a tool for language students to convey a better understanding of Japanese, we also offer the display of detailed information about lexical, syntactic, and transfer knowledge. The system has been implemented in Amzi! Prolog, using the Amzi! Logic Server Visual Basic Module and the Amzi! Logic Server CGI Interface to develop the front ends. Index Terms — natural language processing, machine translation, linguistic knowledge acquisition, parallel corpora, logic programming I.

主題