Objective

study of methods to improve monolingual and cross language access of English and Chinese textual data

Research

• to improve effectiveness of ad-hoc retrieval via web enrichment, heuristics, models
• to predict and improve effectiveness of weak queries
• study of cross language information retrieval including pivot bilingual IR; improve effectiveness via web entity translation
• study of cross language question answering methodologies

Development

• CHINET: a system for English-Chinese named entity translation, cross language name finding, and name extraction from Chinese texts ( http://xpkk.cs.qc.edu:8080/chinet/)
• PIRCS: a multilingual IR system based on an extension of the probabilistic model, supporting non-binary data, query learning and expansion in the style of a neural network