Abstract:
For the second part of the talk, which will be algorithmic, I will revisit the question posed by Nisan and Ronen in the birth of algorithmic mechanism design: How much harder is optimizing an objective over inputs that are furnished by rational agents compared to when the inputs are known? I will present a computationally efficient reduction from mechanism design (i.e. optimizing an arbitrary objective over rational inputs) to algorithm design (i.e. optimizing the same objective over known inputs) in general Bayesian settings. This part of the talk is based on work with Yang Cai and Matt Weinberg.