Exploring the Effectiveness of Chinese-to-English Machine Translation for CLIR Applications in Earthquake Engineering

Abstract

Machine translation is able to support cross-language information retrieval (CLIR) applications given the availability of high-quality translation resources. This study explores the effectiveness of machine translation for the domain of earthquake engineering within the bounds of Chinese-to-English CLIR applications. The writers first defined and then applied a set of procedures to generate Chinese-to-English translation resources for the selected domain. The generated resources are unique and different from general dictionaries or corpora because they address the Chinese segmentation process and are especially oriented for information retrieval (IR) applications. To test the quality of the resources, the writers converted a test collection of documents from Chinese into English through machine translation and conducted trial retrieval runs on the translated documents using different retrieval models. For comparison purposes, the same retrieval runs were also tested on the expert-written English translations. The results suggest that IR-based machine translation was able to yield results comparable to that of human translation and the generated resources had the potential to enable greater sharing of earthquake engineering information across languages while reducing manual translation efforts.