Translating with Scarce Resources

Abstract

Current corpus-based machine translation techniques do not work very well when given scarce linguistic resources. To examine the gap between human and machine translators, we created an experiment in which human beings were asked to translate an unknown language into English on the sole basis of a very small bilingual text. Participants performed quite well, and debriefings revealed a number of valuable strategies. We discuss these strategies and apply some of them to a statistical translation system. Introduction Corpus-based approaches to machine translation (MT) have been on the rise recently, partly because of their promise to automate a great deal of dictionary construction and rule writing, partly because they simply represent a new way of attacking a stubborn problem, and partly because they have performed relatively well in MT evaluations (such as those performed by DARPA and the German Verbmobil program). These approaches generally rely on a large bilingual text cor...