I’ve realized that if I spend as much time reading “research news” as I do reading “tech news”, I’d probably be doing much better in research (in terms of having new ideas, getting inspired with creative thoughts, and just generally to know whats going on elsewhere). So I’ve decided to add a new tag “research” and try and regular post items that are relevant to my research (or just interesting from a research point of view).

So to start off this tag, let me just mention today’s faculty recruit talk. This was a talk by [[http://www-math.mit.edu/~vempala/|Santosh Vempala]]. He’s a faculty in the Math department at MIT, and is now interviewing at some schools for a Computer Science position.

His talk was interesting and impressive in a number of aspects. For one, he did not use powerpoint. Infact, he did not use a computer at all! He did it the old fashioned way — using transparencies and a overhead projector. However, that doesn’t mean his presentation was not good. Quite the opposite — the quality of his slides was exception. Each slide was extremely well thought out, colorful (imagine all the hard work! all slides were done manually) and brought out the relevant points without going into too much detail.

He had a diverse audience, so it was also very nice that he was able to reach out to almost everyone in the audience without losing people in technical details. The talk was on spectral methods and their applications in clustering. The idea was simple, the applications far reaching. To top it all off, he had a [[http://eigencluster.csail.mit.edu/|cool demo]] (try the query ‘jaguar’) and data from some real applications. Works that are based in theory and have some real, pratical applications are the most attractive to me.