Ice-tea: an interactive cross-language search engine with translation enhancement

EXTENDED ABSTRACT In cross-language information retrieval (CLIR), relevance feedback (RF) has been demonstrated to be effective in improving retrieval results, especially when reliable RF information can be obtained from users. Though query expansion (QE) is the leading RF approach in CLIR and it can take place before or/and after translating the query, it should not be the only possible RF method. In our demonstration, besides an implementation of posttranslation QE, we also implement a novel RF approach called translation enhancement (TE) and the integration of TE and QE.