Thursday, November 16, Volen 101, 2:00-3.00 pm
In Game Theory, correlated equilibria generalize Nash ones with significant advantages (easy computability, and higher and flexible payoffs). So far, however, REACHING such equilibria without relying an external trusted party (and indeed, DEFINING what this should mean) has been quite problematic.
We put forward a stronger and more compelling notion of reaching correlated equilibrium, and show that it is always achievable, both efficiently and rationally, via ballots and a ballot-randomizing device.
(Joint Work with Izmalkov, Lepinski, and Shelat)
Host: Liuba Shrira