Optimal Auctions through Deep Learning: Advances in Differentiable Economics
This work addresses the challenge of automated auction design for economists and AI researchers, offering a novel computational approach to a long-standing problem in mechanism design.
The authors tackled the problem of designing optimal auctions for multiple items, which lacks analytical solutions, by using deep learning to automate auction design as a constrained learning problem. They recovered known theoretical solutions and discovered novel mechanisms for unsolved settings through extensive experiments.
Designing an incentive compatible auction that maximizes expected revenue is an intricate task. The single-item case was resolved in a seminal piece of work by Myerson in 1981, but more than 40 years later a full analytical understanding of the optimal design still remains elusive for settings with two or more items. In this work, we initiate the exploration of the use of tools from deep learning for the automated design of optimal auctions. We model an auction as a multi-layer neural network, frame optimal auction design as a constrained learning problem, and show how it can be solved using standard machine learning pipelines. In addition to providing generalization bounds, we present extensive experimental results, recovering essentially all known solutions that come from the theoretical analysis of optimal auction design problems and obtaining novel mechanisms for settings in which the optimal mechanism is unknown.