Detection and translation of OOV terms prior to query time

Accurate cross-language information retrieval requires that query terms be correctly translated. Several new techniques to improve the translation of out of vocabulary terms in English-Chinese cross-language information retrieval have been developed. However, these require queries and a document collection to enable translation disambiguation. Although effective, they involve much processing and searching of the Web at query time, and may not be practical in a production web search engine. In this work, we consider what tasks maybe carried out beforehand, the goal being to reduce the processing required at query time. We have successfully developed new techniques to extract and translate out of vocabulary terms using the Web and add them into a translation dictionary prior to query time.