You are here
Duke-Led Team to Develop Privacy-Minded AI Health Learning Platform
A new $1 million NSF Convergence Accelerator award will support privacy-preserving machine learning models for health care
A whole new world of health care insight is at our fingertips, thanks to quickly evolving machine learning techniques that can analyze vast amounts of data. But how can we effectively leverage this valuable data while ensuring that patient information remains private?
A multidisciplinary research team led by Duke University and composed of Duke electrical and computer engineering faculty members Hai Li and Lawrence Carin, biostatistics and bioinformatics faculty David Page and Erich Huang, and collaborators at the University of Pittsburgh and UMPC (Heng Huang, Wei Chen, Samuel Dickson, Ying Ding, and Liang Zhan) is on track to answer that question. The team has received a new $1 million National Science Foundation (NSF) Convergence Accelerator grant to develop a health model sharing and learning platform named LEARNER.
The NSF Convergence Accelerator program was launched in 2019 to help quickly transition research and discovery aligning with NSF’s “Big Ideas” into practice. In 2020, the NSF continues to invest in two transformative research areas of national importance—quantum technology and artificial intelligence (AI)—to ensure that technological advancements in these areas have a positive impact on society.
“The quantum technology and AI-driven data and model sharing topics were chosen based on community input and identified federal research and development priorities,” said Douglas Maughan, head of the NSF Convergence Accelerator program. “This is the program’s second cohort and we are excited for these teams to use convergence research and innovation-centric fundamentals to accelerate solutions that have a positive societal impact.”
Li is confident that her team’s work will achieve exactly that, by helping to meet the urgent national need for efficient and secure technologies that realize the full potential of big data in health care while addressing FAIR (findable, accessible, interoperable, reusable) data principles and handling the high computation demands of machine learning algorithms. The new platform will also include a repository where data and metadata can be securely collected and shared.
“Our project will fundamentally advance AI-driven health innovations and accelerate use-inspired convergence research in health data science through infrastructure development and framework deployment,” said Li, a machine learning expert who co-directs the Center for Alternative Sustainable and Intelligent Computing (ASIC), an NSF industry–university cooperative research center headquartered at Duke. “The LEARNER platform will support collaborative health data science model sharing and learning, prevent data privacy leakage, and provide complex data and execution management to improve the reproducibility of complex health data analytics.”
Over the next nine months, the 2020 cohort Convergence Accelerator teams will work to develop their initial concept, identify new team members and participate in innovation curriculum focusing on human-centered design, team science, and pitch preparation and presentation coaching. After developing an initial prototype, the teams will participate in a pitch competition and proposal evaluation. Teams selected for phase two will be eligible for additional funding—up to $5 million for a period of 24 months.
By the end of phase two, teams are expected to deliver high-impact solutions that impact societal needs at scale.