An MT Meta-Server for Information Retrieval on WWW

Natural Language Processing for the World Wide Web

Volume

Papers from the 1997 AAAI Spring Symposium

Contents

In the past few years, the World Wide Web (WWW) grows explosively and has become the most useful and powerful information retrieval and accessing system in the Internet. However, the language barrier becomes the major problem to search and retrieve WWW documents in different languages. This paper deals with how to incorporate the capability of language translation into WWW. Several design issues are discussed, including which material is translated, what roles the HTML tags play in translation, the tradeoff between the speed performance and the translation performance, and what form the translated result is presented in. We propose a meta-server architecture to communicate with clients and servers. An online and real-time English-to-Chinese machine translation system is developed. It is acted as a proxy system and an MT service system for retrieving WWW documents. The MT meta-server can not only be used as an assistant tool to access and retrieve information by the end-users in the client, but also can be treated as a Chinese document generating system to produce the Chinese or the bilingual English-Chinese versions of documents dynamically.