Saturday, September 05, 2009

CS and Econ at Cornell: Day I

At Cornell, I got to hear a constellation of Economists and Computer Scientists lay out their research.
  • Larry Blume started it with, let's: Understand better (a) how markets work and should be designed, (b) limitations of Bayesian view, and (c) effects of underlying networks.
  • Sampath Kannan while playing a funding person, still made a scientific argument. He succinctly argued that it was the most natural possible for CS and Econ to bond, both studying human-generated artefacts with Math tools. Historically there should have been more bandwidth in their intersection, and in the future, who knows, may be even a world not of CS or not of Econ in the future, but one of CSECON or ECONCS.
  • Rakesh Vohra did a phenomenal job of representing half a century of interleaved CS and ECON developments, and this in a talk crackling with humor, wit and sharp mind. Later, talking to him, I realized that he can look at the messy world and with a few slices, get to the core and relate it to other worlds.When possible, get him to give a talk or even better, just talk with him.
  • Stephen Morris talked about weakening common knowledge assumptions, and in classic Economics style, showed how revenue is affected when information about others decreases. Jason Hartline spoke exceedingly clearly and with conviction on approximations in mechanism design. He gave nice examples of replacing complex exact designs by simple approximate ones. John Ledyard spoke about Exchanges -- with a cute example of CO_2 exchange eg -- and about fairness, his was the talk I should have given had I been an Economist. Jeff Mackie-Mason spoke about mechanisms to induce online behavior, and how it is difficult to explain humans' altruistic behavior on the web. Michael Kearns spoke about his real-life game experiments. It is a lot of work what they do, and the community should thank them for running these detailed, well-thought out experiments. They should make the data available for the community to analyze. He briefly mentioned how players quickly developed signaling strategies, but not many seemed to have looked for others' signals! Gun Sirer and Jennifer Rexford spoke about CS systems and networking perspectives.
  • Meetings or not, I still had to work, so I had to miss a talk or two.



Post a Comment

<< Home