Andrew Turpin
Professor Andrew Turpin gave a guest seminar on “Query Biased Summaries: Evaluation and Generation” on the 5th July at 11am in room S206 in the engineering building, DCU
Details of the seminar:
In this talk Professor Turpin gave an overview of two projects related to the use of query biased summaries in information retrieval systems that are ongoing at RMIT University in Melbourne. Inevitably, it seems any of our IR projects lead to an exploration of IR system evaluation methodologies. Hence this talk examined how IR systems are compared currently, and how adding summaries to the evaluation can make a marked difference in system comparisons. The second part gave results of a pilot user study that examined how humans generate short, query biased summaries, and how these summaries compare to machine generated summaries.
The talk also included brief sketches of our recent results in stringology. In particular, algorithms for fast pattern matching and string mining.
Biography
Associate Professor Turpin has been the leader of the Information Storage Analysis and Retrieval group at RMIT University, Australia, for the last five years. During that time, his research interests (along with student and staff collaborators) included using web query logs to improve efficiency and effectiveness of search engines, reconciling batch and user IR evaluation methodologies, algorithms and data structures for pattern matching, software authorship attribution, and machine transliteration. He has recently moved to The University of Melbourne to pursue research on computational problems in optometry and ophthalmology, particularly those related to the diagnosis and monitoring of Glaucoma.