I-Corps: Development of machine learning technology for matching under a variety of realistic and largescale preference structures

The broader impact/commercial potential of this I-Corps project focuses on the development of new technologies for achieving efficient two-sided matching under the constraints of fairness and trustworthiness. Numerous technologies presently exist for two-sided matching of supplies and demands. Those assessments that are fair and trustworthy tend to lack efficiency for commercial applications and lack adequate performance specifications, while those with ideal efficiency are too computationally expensive for large-scale markets.

The proposed program explores implementation and commercialization opportunities within the project's initial application focus of small to medium-sized businesses in manufacturing. The proposed technologies have a broad application potential, and can materially reduce the time of finding suitable suppliers and lowering prices. Additionally, the longer-term development of the technology may prove disruptive in markets such as call centers and student tutoring websites.

The developed technologies may also improve the supply chain and better allocate scarce resources, e.g., vaccines and medicines, while ensuring social efficiency and fairness in the matching process.

This I-Corps project is based on the premise that the intersection of state-of-the-art machine learning and economic principles leads to disruptive innovation. The project's technologies offer a fundamentally different approach to two-sided matching than those developed in the past six decades. This project pursues a data-driven solution to a dynamic two-sided matching problem.

Previous progress on this project has focused on the theoretical development of the mechanism design and experimental verification of algorithms in small-scale markets with hundreds of users. By contrast, this project enables a widely scalable approach for millions of users to be matched in real-time and allowing realistic uncertain preferences. The approach not only yields maximum efficiency but also guarantees outcomes such as trustworthiness and fairness.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.