Should we Translate the Documents or the Queries in Cross-language Information Retrieval?

Should we Translate the Documents or the Queries in Cross-language Information Retrieval? J. Scott McCarley IBM T.J. Watson Research Center P.O. Box 218 Yorktown Heights, NY 10598 jsmc@watson.ibm.com Abstract Previous comparisons of document and query translation suffered difficulty due to differing quality of machine translation in these two opposite directions. We avoid this difficulty by training identical statistical translation models for both translation di- rections using the same training data. We in- vestigate information retrieval between En- glish and French, incorporating both trans- lations directions into both document trans- lation and query translation-based informa- tion retrieval, as well as into hybrid sys- tems. We find that hybrids of document and query translation-based systems out- perform query translation systems, even human-quality query translation systems. 1 Introduction Should we translate the documents or the queries in cross-language information re- trieval? The question is more subtle than the implied two alternatives. The need for translation has itself been. questioned : al- though non-translation based methods of cross-language information retrieval (CLIR), such as cognate-matching (Buckley et al., 1998) and cross-language Latent Semantic Indexing (Dumais et al., 1997) have been developed, the most common approaches have involved coupling information retrieval (IR) with machine translation (MT). (For convenience, we refer to dictionary-lookup techniques and interlingua (Diekema et al., 1999) as "translation" even if these tech- niques make no attempt to produce coherent or sensibly-ordered language; this distinction is important in other areas, but a stream of words is adequate for IR.) Translating the documents into the query's language(s) and translating the queries into the docu- ment's language(s) represent two extreme approaches to coupling MT and IR. These two approaches are neither equivalent nor mutually exclusive. They are not equivalent because machine translation is not an invert- ible operation. Query translation and doc- ument translation become equivalent only if each word in one language is translated into a unique word in the other languages. In fact machine translation tends to be a many-to- one mapping in the sense that finer shades of meaner are distinguishable in the original text than in the translated text. This effect is readily observed, for example, by machine translating the translated text back into the original language. These two approaches are not mutually exclusive, either. We find that a hybrid approach combining both directions of translation produces superior performance than either direction alone. Thus our answer to the question posed by the title is both. Several arguments suggest that document translation should be competitive or supe- rior to query translation. First, MT is error-prone. Typical queries are short and may contain key words and phrases only once. When these are translated inappro- priately, the IR engine has no chance to recover. Translating a long document of- fers the MT engine many more opportuni- ties to translate key words and phrases. If only some of these are translated appropri- ately, the IR engine has at least a chance of matching these to query terms. The sec- ond argument is that the tendency of …

Content truncated to 500 words