If we want to improve on something, we have to measure it. We then compare "before" and "after" using some convenient software, baseline system, and measurement metric, and if "after" is bigger/better/brighter than "before", we are done -- we write our paper, and submit it for publication. But behind this apparently simple process are quite complex issues to do with what we measure, how we measure it, and how we report our results. This talk will examine a range of measurement issues in the area of Information Retrieval, including in regard to measuring the efficiency of distributed systems; in regard to caching and other confounding effects; and in regard to the long-standing use of recall as a measure of IR effectiveness.
Professor Alistair Moffat is Head of the Department of Computer Science and Software Engineering at the University of Melbourne, where he has been since he completed his PhD (University of Canterbury) in 1986. Alistair is author or co-author of more than 150 refereed publications in the areas of algorithm design and analysis, text and image compression, and information retrieval; and of three books ("Managing Gigabytes: Compressing and Indexing Documents and Images", 1999; "Compression and Coding Algorithms", 2002; and "Programming, Problem Solving, and Abstraction with C", 2003).
Last modified: Monday, 07-Jul-2008 14:04:58 NZST
This page is maintained by the seminar list administrator .