SIG IR 2006

Search, Life, Tech: I just got back from a whirlwind few days at SIG IR. SIG IR is a rather academic conference on information retrieval that is renowned for having rejected PageRank. Here's a funny little tid bit:

"In a confirmation of what I had heard anecdotally, Brin and Page recount how their research paper on the link ranking feature now known as PageRank was rejected by the SIGIR conference as ‘‘disjointed’’ and ‘‘not IR’’ enough. (It would be interesting to find out who the reviewers were!) In hindsight, one can make a good case that the implementation of Web linking with anchor text was one of the major advances in Web search. Brin and Page’s subsequent paper describing this work (Brin & Page, 1998) has become one of the most cited papers in information retrieval. For those interested, see CiteSeer, ISI Web of Science, and Google Scholar for the citation counts. So, for the legions of researchers who have had papers rejected from SIGIR, take heart!"

Despite this, the sorely lacking keynote speakers, and the monotonous delivery of many a speaker (why oh why can't CS departments across our lovely planet make these PhD students take How-To-Give-A-Speach-101), I still had a great time, and heard some good talks. My favorite was at the closing session by Daniel Frankowski titled You Are What You Say: Privacy Risks of Public Mentions. Basically he showed that its not that hard to link public anonymized users with a private users-not-anonymized data set. This is something especially timely given the AOL public search data release.