Optimal Auctions through Deep Learning

  • 2018-03-19 17:56:32
  • Paul Dütting, Zhe Feng, Harikrishna Narasimhan, David C. Parkes
  • 0

Abstract

Designing an auction that maximizes expected revenue is an intricate task.Indeed, as of today--despite major efforts and impressive progress over thepast few years--only the single-item case is fully understood. In this work, weinitiate the exploration of the use of tools from deep learning on this topic.The design objective is revenue optimal, dominant-strategy incentive compatibleauctions. We show that multi-layer neural networks can learn almost-optimalauctions for settings for which there are analytical solutions, such asMyerson's auction for a single item, Manelli and Vincent's mechanism for asingle bidder with additive preferences over two items, or Yao's auction fortwo additive bidders with binary support distributions and multiple items, evenif no prior knowledge about the form of optimal auctions is encoded in thenetwork and the only feedback during training is revenue and regret. We furthershow how characterization results, even rather implicit ones such as Rochet'scharacterization through induced utilities and their gradients, can beleveraged to obtain more precise fits to the optimal design. We conclude bydemonstrating the potential of deep learning for deriving optimal auctions withhigh revenue for poorly understood problems.

 

Quick Read (beta)

loading the full paper ...