News

Seminar

Impersonation Strategies in Auctions

Dr. Ian Kash, Harvard University

October 06, 2010
Lally 102 - 11:00 a.m. to 12:00 p.m.

Abstract:

A common approach to analyzing repeated auctions, such as sponsored search auctions, is to treat them as complete information games, because it is assumed that, over time, players learn each other's types. This overlooks the possibility that players may impersonate another type. In this talk, I'll show that many standard auctions, including the Kelly mechanism, generalized second price auctions, and core-selecting auctions, have profitable impersonations. I'll define a notion of impersonation-proofness for the process by which players learn about each other's type together with the auction mechanism and associated complete information game and give several examples. Joint work with David Parkes.

Bio:

Ian Kash is a postdoctoral fellow at the Center for Research on Computation and Society at Harvard University working with David Parkes and Yiling Chen on problems at the boundary between economics and computer science such as auctions, kidney exchanges, and prediction markets. He has a B.S. in computer science from Carnegie Mellon University and a Ph.D from Cornell University, where he was advised by Joseph Halpern and Eric Friedman.

Hosted by: Dr. Sanmay Das (x6491)

Last updated: September 28, 2010