no code implementations • 15 Jun 2021 • Michael J. Curry, Uro Lyi, Tom Goldstein, John Dickerson
We propose a new architecture to approximately learn incentive compatible, revenue-maximizing auctions from sampled valuations.