Professor Stephen Robertson

Professor Steve Robertson

Institution: City University, London

Position: Head of Department of Information Science, Director of Centre for Interactive Systems Research

Current research base:
The Centre for Interactive Systems Research consists of three academic staff, five research staff and a number of research students. Most of its work revolves around the Okapi experimental retrieval system, which was originally designed to provide a simple naive-user interface with access to weighting, ranking and relevance feedback (based on a probabilistic model). Okapi runs on Sun workstations, and consists of a retrieval engine and various interfaces for batch or interactive searching, together with some indexing software. The retrieval engine allows Boolean and proximity searching as well as various weighting schemes. Two interactive versions are currently accessible to users outside the Centre, a VT100 interface and a GUI using X-windows. Databases include library catalogues, abstracts databases and full-text datasets such as those used in TREC. Current projects include taking part in the US-based TREC experiment (Okapi produced some of the best results of all the participants in the last round, TREC-3); an experiment combining user query manipulation with relevance-feedback and thesaurus-based expansion in the GUI; and a home-grown database of material from the nineteenth century of interest to literary and cultural historians. Okapi has been designed to allow experimentation with real users in a live-use environment. The Centre has a range of tools for logging live search sessions (on Okapi and other systems), applying online questionnaires where appropriate, and analysing the results.

Experiences:
Robertson is the author of a well-known probabilistic searching model, which has been extensively developed and has proved its value in the TREC experiments noted above. He also has wide experience of retrieval system evaluation, ranging from laboratory-based test collection experiments (like TREC) to live-user experiments. This experience is extended and complemented by that of other members of the Centre, which includes search engine and interface design, and studies of user information-seeking behaviour.

References:
[1] S. E. Robertson and K. Sparck~Jones.
Relevance weighting of search terms. Journal of the American Society for Information Science, 27:129--146, 1976.
[2] S. E. Robertson.
The methodology of information retrieval experiment. In K. Sparck-Jones, editor, Information retrieval experiment, pages 9--31. Butterworths, 1981.
[3] S. E. Robertson and M. M. Hancock-Beaulieu.
On the evaluation of ir systems. Information Processing and Management,, 28:457--466, 1992.
[4] S. E. Robertson, S. Walker, and M. Hancock-Beaulieu.
Large test collection experiments on an operational, interactive system: Okapi at trec. Information Processing and Management, 31:(to appear), 1995.
[5] S. E. Robertson and S. Walker.
Some simple effective approximations to the 2-poisson model for probabilistic weighted retrieval. In W. B. Croft and C. J. van Rijsbergen, editors, SIGIR 94. Proceedings of the 17th International Conference on Research and Development in Information Retrieval, Dublin, pages 232--241. Springer--Verlag, 1994.


Ian Ruthven