New Center to Help Investigators Discover New Knowledge in Medical Databases
July 31, 2007Print-Friendly Version
Delivering adequate health care, preventing illness and developing cures for diseases are increasingly important tasks for a modern society. Progress in these areas, however, critically depends on discovering new medical knowledge.
Ryszard Michalski heads the new Center for Discovery Science and Health Informatics.
Creative Services photo
A highly promising direction toward such a goal is to engage intelligent computers and advanced methods of knowledge discovery to help investigators learn new knowledge from databases that are being accumulated by medical researchers.
Those efforts are being made in the fledgling discipline of health informatics and represent a major research goal undertaken by Mason’s new Center for Discovery Science and Health Informatics. The center is jointly supported by the College of Health and Human Services (CHHS) and the Provost’s Office, and is directed by Ryszard Michalski, PRC Professor of Computational Sciences and Health Informatics.
The new center is dedicated to research on interrelated topics of discovery science and health informatics. Research topics include developing advanced computational theories and methods for data analysis and knowledge discovery and applying them in health care research and practice.
Michalski, a pioneer and cofounder of the field of machine learning, who was recently honored by the president of Poland for his scientific collaboration between scientists in Poland and the United States, notes that the creation of the center will contribute to education and research in CHHS. It will also enable health research faculty, as well as other Mason researchers, to have easy access to the newest methods for knowledge discovery from data and prior knowledge supplied to the computer.
The center is composed of two closely related units: the Machine Learning and Inference Laboratory (MLI) and the Health Services Research Unit (HSR). MLI, an already existing laboratory at Mason, is concerned with the development of algorithms and systems that allow computers to “learn.” These systems enable computers to extract rules and patterns out of large and small data sets. HSR adapts and applies the developed methods to problems of medical research and health care delivery.
“Combining the two units into one center will facilitate ongoing research collaborations, and will open doors for new opportunities for collaborations in the area of health informatics, both within Mason and outside organizations, such as the National Institutes of Health,” says Michalski.
A hallmark of the center’s research is that it develops and applies methods for creating knowledge in forms that are easy to understand and interpret, which is a particularly important requirement in medicine. Such forms include natural language descriptions and easy-to-interpret graphical representations. The methods integrate results from machine learning, artificial intelligence, software engineering and cognitive science.
According to Shirley Travis, dean of CHHS, the center will aid physicians, clinicians and researchers, such as Lynn Gerber, director of the Center for the Study of Chronic Illness and Disability in CHHS. Gerber, for example, could take the information she has collected about fatigue, enter it into a computer and discover how it might be related to chronic illness. A collaboration on this topic is already ongoing.
“We’re hoping that as a result of our reaching out to other physicians and clinicians in the community they will bring their research to the center and use the computers to develop a way to make a decision on medical problems,” says Travis. “Perhaps the center will help lead researchers to discover new knowledge about diseases such as Alzheimer’s.”
Currently, Michalski and his colleagues are also working with Ancha Baranova, Mason professor in the Department of Molecular and Microbiology in the College of Science; and Zobair Younossi, executive director of research and the Center for Liver Diseases at Northern Virginia’s Inova Health System, on problems of discovering patterns in patients with metabolic syndrome.
For more information about the Center for Discovery and Health Informatics, visit the web site.