I-Corps: Privacy-preserving data sharing software platform

University Of California-Santa Barbara, Santa Barbara CA

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project is the development of a data sharing product for enterprises in highly regulated verticals such as healthcare, finance, and the government. Today, data custodians in these verticals must keep their confidential data assets in silos due to risks of data breaches, regulatory violations, and reputational setbacks. The proposed technology may enable these enterprises to build a centralized database and expose this data safely for the application of artificial intelligence and machine learning, leading to direct economic and societal benefit. As an example, in the government sector, the proposed technology may enable different departments of Health and Human Services, Judicial System, and Community Services to understand risk factors for incarcerations such as homelessness and serious mental illness, to better allocate resources, design effective interventions, and reduce social inequities. The challenge in building the centralized database is security and privacy, and the project will bring to the forefront the very best of security and privacy technologies. This I-Corps project is based on the development of a privacy-preserving data warehouse and analytics engine. This warehouse connects to heterogeneous and federated data sources, while ensuring that the original confidential data stays at source with the data custodians. The data warehouse links records from different systems of records while removing duplicates in a privacy-preserving manner. The secure analytics engine makes the linked data available for queries, while providing anonymity guarantees through the state-of-the-art technology of differential privacy. The query results obtained do not disclose the identity of an individual whose record is present in the data. Based on fundamental research, the data warehouse and analytics engine include novel system architectures for federated queries, design optimizations for performance and scalability to big data workloads, and rigorous algorithms for anonymity. 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.