Abstract
Prior-independent mechanism design asks for a single mechanism, chosen without knowing the distribution of agents’ preferences, that nevertheless competes with the Bayesian optimal mechanism for every distribution in a specified class. This talk surveys the past, present, and future of this robustness framework. We begin with the canonical two-agent, single-item auction setting, where the optimal prior-independent mechanism is a randomized markup mechanism achieving approximation ratio about 1.91. We then discuss a lower-bound framework based on blend pairs, which exposes the minimax structure behind robust mechanism design. Next, we connect prior-independent optimization to prior-free benchmark design, showing that optimized normalized benchmarks and optimal prior-independent mechanisms are two views of the same problem. Finally, we turn to a more subtle issue: prior independence breaks the usual revelation-principle reduction, opening the door to revelation gaps in which non-truthful mechanisms outperform truthful ones.0Speaker
Jason Hartline is a professor of computer science at Northwestern University. Prof. Hartline received his Ph.D. in 2003 from the University of Washington under the supervision of Anna Karlin. He was a postdoctoral fellow at Carnegie Mellon University under the supervision of Avrim Blum; and subsequently a researcher at Microsoft Research in Silicon Valley. He joined Northwestern University in 2008. He was on sabbatical at Harvard University in the Economics Department during the 2014 calendar year and visiting Microsoft Research, New England for the Spring of 2015. He was on sabbatical at Stanford University for the 2023-2024 academic year.
Prof. Hartline is the director of Northwestern’s Online Markets Lab, he was a founding codirector of the Institute for Data, Econometrics, Algorithms, and Learning from 2019-2022, and is a cofounder of virtual conference organizing platform Virtual Chair.
Room
Room 104
