This project seeks to provide a novel solution for explainable prescriptive analytics that draws collective intelligence of experts to optimize enterprise performance management (EPM) decisions. The project will develop a system named Knowledge Graph-Based Reasoning Artificial Intelligence Network (KG-BRAIN). KG-BRAIN proposes to help organizations deal with complex EPM tasks concerning multiple stakeholders’ demands. Leveraging a partnership of researchers and commercialization experts among two public universities, a large-scale healthcare performance improvement platform, a seasoned consulting firm on EPM, and the region’s largest incubator, the team will develop a commercial demonstration for healthcare systems, generalizable to other industries.

Given various stakeholders’ demands, EPM is a complex and challenging task for all organizations. For example, healthcare systems need to improve the quality of patient care and outcomes, drive clinician satisfaction, and increase resource use efficiency and deliver returns to investors. Most healthcare systems in the U.S. struggle to maintain their performance, with a 21.3% decline in operating margins before COVID-19 from 2018 to 2019; nearly 50% of rural hospitals are operating at a loss. The COVID-19 crisis further exacerbated the performance management challenges of the U.S. healthcare systems, with an estimated $50.7 billion in losses per month during the pandemic. Meanwhile, only 54% and 30% of Americans in 2020 were satisfied with the quality and cost of U.S. healthcare, respectively.

In a recent NSF I-Corps (NSF Award #2102803), the project team interviewed 145 leaders in EPM, more than 90% of whom suggested an unaddressed need for explainable prescriptive analytics on EPM. The key technical hurdle is the increasing gap between data science and subject matter expertise. One persistent challenge for effective EPM is the lack of an enterprise-wide, integrated analytics system to prescribe optimal management actions. A 2020 Gartner survey shows that more than 50% of healthcare systems worldwide are not satisfied with the EPM analytics software, quoting that the most critical impediment for long-range planning is fragmented insights.

This project seeks to bridge this gap with an explainable and causal AI system by translating and integrating the team’s prior research in EPM knowledge graph (NSF Award #2102803), machine reading (NSF Award #2006583), and visual informatics (NSF Award #1747785). KG-BRAIN will (a) automatically synthesize fragmented causal statements from both practitioners and the literature in social, behavioral, and economic sciences, (b) visually represent synthesized knowledge as causal knowledge graphs, and (c) intuitively run causal data analysis.

Principal Investigator:

Dr. Victor Zitian Chen - UNC Charlotte



Co-Principal Investigators:

Dr. John Martin - Premier Inc.

Dr. Gus Hahn-Powell - University of Arizona

Dr. Wenwen Dou - UNC Charlotte

Dr. David J. Woehr - UNC Charlotte

Senior Personnel:

Dr. Wlodek Zadrozny - UNC Charlotte

Dr. Reginald Silver - UNC Charlotte

Mike Korvink - Premier Inc.

Gary Cokins - Analytics-based Performance Management LLC (Industry Mentor)

Dan Roselli - RevTech Labs (Technology Commercialization Specialist)

Key Technological Advantages of KG-BRAIN:

• Three prior NSF funded projects
• 5 granted and 1 pending patents
• A unique interdisciplinary integration of multiple decades' cutting-edge research in management, knowledge synthesis, machine reading comprehension, and visualization

Key Commercial Advantages of KG-BRAIN:

• Partnership with one of America's largest healthcare platforms serving 4,100+ hospitals and 200,000 other healthcare organizations
• Partnership with seasoned industry mentor with 40+ years' executive and consulting experience in EPM analytics
• Partnership with the region's oldest and largest incubator, whose alumni impacts include $2.5+ billion in VC and $300 million in strategic exits such as IPOs