Mechanism-Informed Learning for Fair Division

Ayumi Igarashi, Ryota Maruo, Koh Takeuchi, Tomohiko Yokoyama

Keywords: Fair division, Machine learning

TL;DR: We proposed a learning framework for fair allocations from incomplete preferences.

Abstract: Fair division provides a simple yet powerful framework for modeling fairness in resource allocation. While existing literature typically assumes complete information about preferences, many practical scenarios involve incomplete preference data, posing challenges in estimating fair allocations. In this paper, we propose mechanism-informed preference learning, a framework that integrates neural networks with differentiable approximations of classical fair division mechanisms---adjusted winner, round-robin, and moving-knife---to estimate fair allocations from incomplete preference. Experiments on real-world household chore preference data show that our approach using differentiable mechanism achieves fairer allocations, compared to methods that do not incorporate mechanism information.

Area: Game Theory and Economic Paradigms (GTEP)

Generative A I: I acknowledge that I have read and will follow this policy.

Submission Number: 1031