People: Stephen Baek, assistant professor of industrial and systems engineering
Mission Area: Biotechnology
Lab: Visual Intelligence Laboratory
The National Science Foundation (NSF) awarded a $1 million phase one grant earlier this month to an Iowa-led, multi-university consortium to advance medical imaging research through artificial intelligence (AI).
The research team is part of the AI-driven data and model sharing track topic under the 2020 cohort NSF Convergence Accelerator program designed to leverage a convergence approach to transition basic research and discovery into practice. University of Iowa Technology Institute (ITI) faculty affiliate Stephen Baek leads the grant as the principal investigator. ITI faculty affiliate Xiaodong Wu, professor of electrical and computer engineering, and Nick Street, professor of business analytics in the Tippie College of Business, are also investigators on the grant.
Collaborators from Stanford University, University of Chicago, Harvard University, Yale University, and Seoul National University, as well as industry medical AI leaders from NVIDIA, Lunit, Digital Diagnostics (formerly known as IDx Technologies), and inSEER will also be part of the consortium.
"Despite the recent breakthroughs in AI/ML (machine learning) for medical imaging, these achievements have not yet yielded tangible benefits to general clinics," Baek says. "The breakthroughs in AI have generated considerably heightened expectation towards the next generation patient care. In actuality, however, the constituents of the breakthroughs so far have mostly been individual research studies administered by a single, or small group of, institutions, due to administrative, technical, and regulatory restrictions regarding patient data sharing."
The group plans to "develop a novel federated learning (FL) method to overcome such limitations in patient data sharing," says Baek, an assistant professor of industrial and systems engineering and director of the Visual Intelligence Laboratory.
A focal point of the work will be to design and develop a usage-inspired software platform, called ImagiQ, centered around the idea of asynchronous and decentralized federated learning.
“ImagiQ will further federated learning by decentralizing the model training,” said Baek. “We are going to create a whole ecosystem of machine learning models that will evolve and improve over time. ImagiQ will build a better AI system for analyzing medical images by making diverse medical image data available for training without actual data sharing.”
Over the next nine months, the team will focus on making a prototype of the system as well as participate in the Accelerator’s innovation curriculum to ensure the solution has societal impact. At the end of phase one, the team will participate in a pitch competition and a proposal evaluation and, if selected, will proceed to phase two, with potential funding up to $5 million for 24 months.
Here's more about the Convergence Accelerator program from an NSF announcement:
The U.S. National Science Foundation has selected 29 teams for its Convergence Accelerator program, a new NSF initiative designed to accelerate use-inspired research to address wide-scale societal challenges. The 2020 cohort addresses two transformative research areas of national importance: quantum technology and artificial intelligence.
NSF is investing more than $27 million to support the teams in phase one to develop the solution groundwork for Quantum Technology, and AI-Driven Data and Model Sharing to ensure that technological advancements have a positive impact on society.
Over the next nine months, the teams will work to build a proof-of-concept for their solutions by using fundamentals from the Convergence Accelerator's model and curriculum such as: multidisciplinary partnerships between academia, non-profits, government, industry, and other sectors; human-centered design thinking; team science; early-stage prototyping; pitch preparation; and use-inspired research, which engages the end-user early in the design process to ensure the solutions address issues of significant national impact.