UB announces first round of seed funding for health projects integrating AI

BUFFALO, N.Y. — Drugs customized to a person’s DNA. Streamlined hospital admissions for older adults with complex medical issues. Improved language development in children who are late talkers. A way to enhance surgical skills and patient outcomes.

Now underway at the University at Buffalo, these four projects spanning multiple disciplines are using artificial intelligence to enhance health care. Each was awarded $50,000 in UB’s first round of competitive interdisciplinary seed funding for AI research in health care.

The funding is being provided through a collaboration between the UB Office of the Vice President for Research and Economic Development and the Office of the Vice President for Health Sciences.

“With UB as the home of Empire AI, our Institute for Artificial Intelligence and Data Science, and our six health sciences schools plus engineering, UB clearly has the talent and experience to lead AI in health,” says Venu Govindaraju, PhD, vice president for research and economic development at UB. “As these four projects demonstrate, our researchers are actively innovating across all disciplines to leverage AI for the common good.”

Announced earlier this year by Gov. Kathy Hochul and approved by state lawmakers, Empire AI is a consortium of public and private higher education and philanthropic partners across the state. It aims to secure New York’s place at the forefront of artificial intelligence research with the goal of accelerating research and innovation in AI. The consortium’s computing center, to be located at UB, will be used by leading New York institutions to promote responsible research and development, create jobs and advance AI for the public good.

“As medical professionals, we are enthusiastic about harnessing the power of AI to tackle some of society’s most urgent health challenges,” says Allison Brashear, MD, vice president for health sciences and dean of the Jacobs School of Medicine and Biomedical Sciences at UB. “The creativity and innovation these researchers have demonstrated reflects the potential of AI to bring truly game-changing innovation to the clinic and the bedside to provide tangible benefit to patients and their caregivers. Their cross-disciplinary approaches harness the many strengths of UB and will bring forward creative and impactful solutions.”

UB’s call for AI in health research proposals generated more than 40 applications from the more than 200 researchers at UB who are working on AI innovations in drug discovery, medicine, robotics and throughout health care and beyond. A requirement for submissions was that proposals needed to include faculty members from different academic units to ensure the teams reflected cross-disciplinary expertise.

Of the 40 submissions, 12 research proposals were presented at UB’s first research in AI and health care symposium last winter, where they received feedback from UB leadership. Mary Ellen Giger, PhD, A.N. Pritzker Distinguished Service Professor of Radiology at the University of Chicago, gave the keynote talk.

The goal of this first round of pilot funding is to provide UB researchers with the opportunity to generate preliminary results and eventually attract more significant funding from the National Institutes of Health and other federal agencies. Seed funding for projects like these is essential to allow researchers to collect preliminary data, especially in emerging fields.

The four teams will present their preliminary findings at the next health and AI symposium sponsored by the Office of the Vice President for Health Sciences on Oct. 4, when David C. Rhew, MD, global chief medical officer and vice president of health care for Microsoft, will be the guest speaker.

The principal investigators describe their projects below:

• SWAXSFold: A New AI Tool to Determine Protein Structures

Thomas D. Grant, PhD, assistant professor of structural biology, Jacobs School.

“SWAXSFold brings together the latest advances in AI modeling with powerful experimental X-ray scattering to model protein structures with higher accuracy than ever before,” Grant says. “In the future, this will enable a new generation of targeted, precision medicine, where drugs are designed for individual patients based on their own DNA.”

• AI to Identify Protective Factors for Children with Delayed Language Development

Federica Bulgarelli, PhD, assistant professor of psychology, College of Arts and Sciences

“In this project, we are hoping to test whether some aspects of what parents say to kids can serve as naturally protective factors for later language delay diagnoses,” says Bulgarelli. “If we are able to identify these potentially protective factors, not only may we be able to identify children who are at risk earlier, but we may also be able to encourage parents to shift their language to focus more on these protective properties of speech.”

• Leveraging AI for Clinical Summaries in Hospital Care

Sabrina Cascucci, PhD, assistant professor of industrial and systems engineering, School of Engineering and Applied Sciences.

“Our team is extremely excited to develop new AI methods and health information exchange solutions for generating care summaries from longitudinal, multisource, multimodal historical care records at the time of hospital admission,” says Cascucci. “This is a critical challenge in the care of complex older adults as developing an accurate and meaningful understanding of community-based care and understanding how these factors impact readmission risk can have a significant impact on hospital-based care and post-hospital health outcomes for these vulnerable patients.”

• SurgiVdoNet: A Digital Common for Surgical Videos

Gene Yang, MD, clinical assistant professor of surgery, Jacobs School.

“SurgiVdoNet is a digital common for annotated surgical video providing a data repository for ethical surgical artificial intelligence development,” Yang says. “Blockchain ledger technology provides secure and robust tracking to provide permanent ownership and attribution for all stakeholders including patients, hospitals and physicians. SurgiVdoNet provides a rich source of high quality, annotated surgical video data for training, testing and development of state-of-the-art, generalizable AI models to democratize surgical skill and improve patient outcomes.”