Ease of access to foreign sites via the web has led to recent popularity in CLIR research. Resources like large multi- lingual dictionaries, linguistic tools, etc. in general are of primary importance for such work. I will describe our investigation into English-Chinese CLIR with the TREC Chinese collections of general domain. Low level resources were used: an affordable translation software and an available bilingual dictionary. Results show that we can achieve over 70% of monolingual effectiveness.
Retrieval is based on our PIRCS bilingual IR system. This system has been highly successful in past TREC blind experiments. The characteristics of PIRCS will be described as well as the techniques used for CLIR.
For the colloquium series schedule, see the UMD Computational Linguistics Colloquium Series web page at http://umiacs.umd.edu/~resnik/cl_colloquium/. If you are interested in meeting with the speaker, please contact Philip Resnik (resnik@umiacs.umd.edu).