Learning Potentials for Dynamic Matching and Application to Heart Transplantation

2026-02-09

AI Summary

Current heart transplant allocation policies suffer from suboptimal population-level efficiency due to their inability to effectively respond to the dynamic arrival of donor organs and evolving characteristics of waitlisted candidates. This work proposes a far-sighted online matching framework grounded in potential-based decision-making, introducing for the first time a high-dimensional, highly expressive potential function trained via self-supervised imitation learning to approximate the oracle policy with perfect foresight. By striking a balance between theoretical rigor and computational scalability, the proposed method significantly outperforms both the current U.S. allocation policy and state-of-the-art continuous allocation frameworks on real-world heart transplant data, yielding substantial improvements in key outcomes such as overall patient survival rates.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Online Learning & Bandits
• Humans and AI: Learning Human Values and Preferences

Application Category

• Economics, Online Markets and Human Computation: The sharing economy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Each year, thousands of patients in need of heart transplants face life-threatening wait times due to organ scarcity. While allocation policies aim to maximize population-level outcomes, current approaches often fail to account for the dynamic arrival of organs and the composition of waitlisted candidates, thereby hampering efficiency. The United States is transitioning from rigid, rule-based allocation to more flexible data-driven models. In this paper, we propose a novel framework for non-myopic policy optimization in general online matching relying on potentials, a concept originally introduced for kidney exchange. We develop scalable and accurate ways of learning potentials that are higher-dimensional and more expressive than prior approaches. Our approach is a form of self-supervised imitation learning: the potentials are trained to mimic an omniscient algorithm that has perfect foresight. We focus on the application of heart transplant allocation and demonstrate, using real historical data, that our policies significantly outperform prior approaches -- including the current US status quo policy and the proposed continuous distribution framework -- in optimizing for population-level outcomes. Our analysis and methods come at a pivotal moment in US policy, as the current heart transplant allocation system is under review. We propose a scalable and theoretically grounded path toward more effective organ allocation.

Problem

• dynamic matching
• organ allocation
• heart transplantation
• online matching
• policy optimization

Innovation

• potentials
• non-myopic policy optimization
• online matching
• self-supervised imitation learning
• dynamic organ allocation