Tim Tai, Photography Editor

Yale’s newest data science initiative, announced in August, has opened its doors.

The Institute for Foundations of Data Science, otherwise known as both the Kline Tower Institute and as FDS, launched on Oct. 14 as part of an ongoing effort by the University Science Strategy Committee, who identified the integration of data science and its mathematical foundations on campus as its top priority. The kickoff event featured more than 20 faculty members who briefly presented their own research in areas of data science.

“What our institute is doing is really more about developing core data science, as opposed to thinking about applying it within specific disciplines,” said Sterling Professor of computer science and professor of statistics and data science and mathematics Daniel Spielman, the inaugural director of FDS. “These [foundations] are the things that we expect to be useful across disciplines and help us understand how data science works.”

Vice Provost for Research Michael Crair also emphasized the importance of understanding large and complex data, stating that the problems and solutions spanning across academic fields can have similar underlying mathematical structure.

Spielman added that FDS is committed to supporting faculty from the different departments that conduct a lot of data-driven research, including the Departments of Statistics and Data Science, Computer Science, Electrical Engineering, Mathematics, Economics, Political Science and even Biostatistics at the Yale School of Medicine.

Spielman noted that by bringing the University’s data science community together, the institute will inspire more conversations between people from various backgrounds and will lead to the increase of interdisciplinary research.

“For a good research project to happen, you need people to meet,” Spielman said.

He clarified that FDS will serve the role of bringing together “people who have compatible interests or [who] can merge together to do something useful.”

According to Rohan Khera, assistant professor of medicine and biostatistics, the initiative will provide postdoctoral research training in an interdisciplinary setting, as well as weekly seminars led by mentors and faculty members that will showcase the wide range of applications of data science.

Khera shared his enthusiasm for a spring workshop on healthcare for data scientists that he and Spielman are currently working on together.

“We will invite a lot of different thought leaders who have experience in doing healthcare and data science,” Khera said.

Director of Undergraduate Studies in biomedical engineering James S. Duncan also acknowledged the growing domain of data science, saying he looks forward to working with people who develop similar methods as him in both related and disparate disciplines.

Duncan’s current research focuses on functional magnetic resonance imaging, or fMRI, of the brains of children with autism in order to understand what parts of the brain are involved with this condition and be able to better predict the outcome of a number of different cognitive therapies.

Duncan explained that his work requires the incorporation of various machine learning strategies that are applicable to other fields such as genetics and astronomy.

“One of my goals is to tap in and see what others are doing in those areas and what their ideas are and how it relates to the problems I’m working on,” Duncan said.

Professor of earth and planetary sciences Jeffrey Park, whose work is focused on time series analysis and spectrum estimation in the geosciences, echoed Duncan’s sentiment. Park said that he hopes to gain insight into his research areas by learning from experts in other fields.

Park added that having a large room where people see each other’s works being presented is “what the Yale experience should be for both researchers and students.”

Data science is also a critical component of the research done by Amin Karbasi, associate professor of electrical engineering and computer science. Karbasi, who led the first seminar on Sept. 21 called “Gaming the Learning,” shared that his talk focused on data labeling, a topic he believes is often overlooked despite its role in developing efficient machine learning and artificial intelligence methods.

FDS is expected to move to Kline Tower in August 2023 after its renovation is complete, along with the Faculty of Arts and Sciences Departments of Astronomy Mathematics and Statistics and Data Science. Kline Tower will also welcome members of the FAS Department of Physics.

According to Karbasi, the opportunity to have a physical space is “crucial,” as interactions between researchers across disciplines would facilitate discussions that would not otherwise happen.

“Real science comes together when people meet and discuss in long hours,” Karbasi said.

Spielman added that spaces like Kline Tower help build a community of scientists and offer a physical location where the “data science party” can happen. 

Kline Tower is currently the tallest building on campus.