GemNet: Menu-Based, Strategy-Proof Multi-Bidder Auctions Through Deep Learning

TL;DR

GemNet is a deep learning-based approach that designs strategy-proof multi-bidder auctions with improved revenue, exact strategy-proofness, and interpretability, extending menu-based auction design to complex multi-bidder settings.

Contribution

It introduces a novel deep learning framework for fully strategy-proof multi-bidder auctions using menu compatibility and transforms, advancing automated mechanism design.

Findings

Achieves exact strategy-proofness in multi-bidder auctions.

Outperforms affine maximization methods in revenue.

Scales to large auction problems with deep learning optimizations.

Abstract

Automated mechanism design (AMD) uses computational methods for mechanism design. Differentiable economics is a form of AMD that uses deep learning to learn mechanism designs and has enabled strong progress in AMD in recent years. Nevertheless, a major open problem has been to learn multi-bidder, general, and fully strategy-proof (SP) auctions. We introduce GEneral Menu-based NETwork (GemNet), which significantly extends the menu-based approach of the single-bidder RochetNet (D\"utting et al., 2024) to the multi-bidder setting. The challenge in achieving SP is to learn bidder-independent menus that are feasible, so that the optimal menu choices for each bidder do not over-allocate items when taken together (we call this menu compatibility). GemNet penalizes the failure of menu compatibility during training, and transforms learned menus after training through price changes, by considering a set of discretized bidder values and reasoning about Lipschitz smoothness to guarantee menu compatibility on the entire value space. This approach is general, leaving trained menus that already satisfy menu compatibility undisturbed and reducing to RochetNet for a single bidder. Mixed-integer linear programs are used for menu transforms, and through a number of optimizations enabled by deep learning, including adaptive grids and methods to skip menu elements, we scale to large auction design problems. GemNet learns auctions with better revenue than affine maximization methods, achieves exact SP whereas previous general multi-bidder methods are approximately SP, and offers greatly enhanced interpretability.