Cross-Language Multi-Media Information Retrieval

We describe here the evaluation of a cross-language information retrieval tech-nique based on similarity thesauri in a multi-media document environment, such as is likely to be found in a digital library. We present the theory of simi-larity thesauri, which are information structures derived from corpora, and show how they can be used for cross-language retrieval. In evaluating our similarity thesaurus-based approach to cross-language retrieval over a parallel collection of legal texts, we show that cross-language retrieval can perform equally as well as monolingual retrieval in the certain cases. We also present the results of a rst evaluation of cross-language retrieval with spoken news material. We conclude that providing cross-language access to multi-media digital libraries is already a viable possibility. 1