University announces $100 million data science initiative

Tuesday, September 8, 2015 - 5:28pm



The University will invest $100 million in a new Data Science Initiative over the next five years with the aim of enhancing learning and research opportunities for students and faculty members.

To support the initiative, the University will hire 35 new faculty members over the next four years and launch the Michigan Institute for Data Science, which will lead educational and research opportunities related to big data. Massive sets of data can help researchers produce new insights into a broad spectrum of topics, from learning and medicine to transportation and social media.

"Big data can provide dramatic insights into the nature of disease, climate change, social behavior, business and economics, engineering, and the basic biological and physical sciences," University President Mark Schlissel wrote in a statement. "With our widely recognized strengths across all of these areas and our longstanding culture of collaboration across disciplines, U-M is in a unique position to leverage this investment in data science for the good of society."

The initiative also aims to expand the University’s computing capacity, support interdisciplinary research on big data and provide opportunities for students interested in careers related to data science.

Engineering Prof. Alfred Hero and Brian Athey, a professor of psychiatry and internal medicine who chairs the Department of Computational Medicine and Bioinformatics, co-direct MIDAS.

They said multidisciplinary institutions like the University fuel researching by using data from very different domains. Discoveries in archeology, Hero said, are driven not only by findings by archaeologists at dig sites, but also by archived historical data and language processing machines. The big data MIDAS is concerned with analyzing, Hero and Athey said, has to do with its variety and velocity.

“The idea about big data is not just about the volume or the size or how many bytes it is,” Athey said. “(It’s about) how complex and diverse the data is. It’s about variety … velocity. How fast is it growing? We anticipate there could be terabytes streaming in a minute. It could be that fast.”

The DSI will focus on collecting and using big data to strengthen transportation research, health sciences, learning analytics and social science research.

“Big data is revolutionizing research in extraordinary range of disciplines,” S. Jack Hu, interim vice president for research, wrote in a statement. “With this initiative, our goal is to spark innovation in research across campus while inspiring further advances in the techniques of data science itself.”

Using patient visits, medical histories and DNA sequences, for example, researchers can access an individual’s risk for certain types of cancer.

The University also hopes to use student data to customize teaching to the needs of individual students.

Data security will be an important part of the DSI, Hero said.

“There is a classic paradigm between data security and data privacy,” Hero said. “That paradigm really starts to break down with big data. Despite all of your efforts to protect your own personal data, it can be eventually used to triangulate down to geography and, in some cases, identity can be compromised. Privacy means more than just protection of your data by the standard identification techniques. It means you have to develop new methods.”