CS&E alums and Professors Konstan and Riedl receive the SIGIR Test of Time Award

CS&E alumni Jon Herlocker and Al Borchers, along with Professors Joe Konstan and John Riedl (posthumous), have been recognized with the ACM’s Special Interest Group on Information Retrieval’s (SIGIR) Test of Time Award. They were honored for their 1999 paper, “An algorithmic framework for performing collaborative filtering.

Their paper, published just five years after the first work on automated collaborative filtering systems, introduced a rigorous and scientific approach to developing and tuning those systems.  Collaborative filtering systems—an important class of recommender systems—collect ratings from people for items and match them with people who share the same information needs or taste.  For example, Amazon.com and Netflix both use collaborative filtering, basing recommendations for products and movies on a customer's prior selections and the other customers who have similar tastes.  

In the early days, builders of such systems were happy simply to see the systems work at all.  While each system was based on the same core ideas, they were implemented in a widely varying set of ways.  This paper took a systematic approach to cataloging the variations and then experimentally comparing them and measuring the impact on three key factors—how accurately the system could predict preferences, how well the system distinguished items a customer would like from those they wouldn't, and how broadly the algorithm could cover the entire product catalog.  This approach of evaluating algorithms on the basis measures of customer-experience has remained an important part of the field of recommender systems.  

The SIGIR Test of Time Award recognizes research that has had lasting influence across and beyond the research community involved with information retrieval.  A special awards session was held to honor the GroupLens team and other honorees at The 40th International ACM SIGIR Conference on Research and Development in Information Retrieval that took place in Tokyo, Japan.  The original article was reprinted in a special issue of SIGIR Forum.

Please join CS&E in congratulating the entire team on receiving this award.