Connecting Data Science to ‘Almost Every Domain of Inquiry’

If you don’t believe that big data and artificial intelligence are here to stay, just ask Alexa: “What academic disciplines this fall are driving two major research universities to reorganize?”

The University of California, Berkeley, and Massachusetts Institute of Technology are doing just that, creating entirely new institutions within their campuses to come to terms with the ubiquity of data and the rise of AI -- and to accommodate a surge in popularity that these fields are generating among students and employers.

Berkeley on Thursday said it plans to form an entirely new division, to be tentatively called the Division of Data Science and Information, that will engage faculty members and students across the flagship UC campus. The division, officials said, will be led by a new associate provost and connect departments as disparate as UC’s College of Engineering and its College of Letters and Science.

The announcement follows a similar one last month at MIT, which unveiled plans to build an entirely new, $1 billion college devoted to AI, data science and related fields. MIT said it had already received a “foundational gift” of $350 million from Blackstone co-founder Stephen A. Schwarzman. The new college, on MIT’s Cambridge, Mass., campus, will bear his name.

MIT Schwarzman College of Computing will create 50 new faculty positions, both within its walls and across other departments, MIT said. It will “reorient MIT to bring the power of computing and AI to all fields of study,” nearly doubling the institute's academic computing staff and giving its five schools “a shared structure for collaborative education, research and innovation in computing and AI.”

Marty Schmidt, MIT’s provost, has said 40 percent of MIT undergraduates now major either in computer science or earn a joint degree that includes computer science. Teaching all of those students puts a strain on MIT’s existing computer science faculty, he said.

But these students "don’t want to code for life," Schmidt said. "They want to gain these skills that allow them to apply them in their discipline."

To that end, MIT also is pushing faculty members to become what it calls “bilinguals” who have a foot both in their discipline and in computation. “That creates individuals who can really bring these advanced tools to the disciplines,” he said.

MIT previously has created interdisciplinary research labs but struggled to bring that approach to instruction, Schmidt said. To make it happen on a large scale, MIT decided it needed a new college, not just an expanded computer science department.

Berkeley provost Paul Alivisatos said that simply expanding the university's existing computer sciences department would not be enough to match the surge of interest.

“Pretty much any field of inquiry and knowledge connects to [data science],” he said. “We wanted to create a structure that would allow that new methodological development to grow more, but also allow it to be widely used everywhere, where it can be beneficial.”

He said Berkeley envisions incorporating faculty members from fields as varied as sociology, public health and physics into a kind of “data science commons” to deepen their research. “From what we can tell, pretty much every part of this university wants to be involved, which is great.”

The field, Alivisatos said, is forcing other disciplines to come to terms not just with the widespread availability of data from diverse sources, but with “new methods that allow it to be sifted and analyzed.”

David Culler, Berkeley’s interim dean for data sciences, said the new division will be a peer of the university’s other schools and colleges. “But rather than standing apart from them, it’s really integrated with them,” he said, since these days, data science “touches almost every domain of inquiry.”

Culler said Berkeley, like most major universities, has been “grappling with this for at least five years” as it tried to figure out how to fit new computational disciplines into the broader world of other academic fields.

“The frontiers of knowledge are extremely integrative, and yet to a large extent, institutions of higher learning are very hierarchical,” he said.

Berkeley officials said another reason for the new division has been the sheer popularity of data science among students: two recent data science courses have turned out to be among the fastest-growing ever, they said.

One of them, Foundations of Data Science, saw enrollment grow from 100 in 2015 to 1,300 this fall. Its more advanced follow-up course, Principles and Techniques of Data Science, grew from 100 students in 2016 to 800 this fall. Berkeley this fall also announced a new integrative data science major.

“The moment we opened up the major, we started getting a flood of predeclarations,” Alivisatos said, noting that 1,070 students have already said they plan to major in data science after their first year of preparatory course work.

Alivisatos said he was talking last Monday with a biology student who told him he’d taken the data course and decided to use what he’d learned to more closely analyze data sets on bird flight. Another, who is earning a Ph.D. in law, said he'd become convinced that big data represents the future of the legal profession.

“My sense is that this is not a casual transformation,” he said. “This is something that’s going to be here a very long time.”

Recent research has suggested that a shortage of job candidates with fluency in data science and analytics is among the nation’s most yawning of skills gaps. The Business-Higher Education Forum, a nonprofit membership group of Fortune 500 CEOs, college leaders and the consulting firm PricewaterhouseCoopers, found last year that by 2020, about 2.72 million new job postings will seek workers with skills in data science and analytics.

Cathy O’Neil, a onetime MIT postdoc and hedge fund "quant" who launched a data journalism program at Columbia University's Graduate School of Journalism in 2014, said the expanded academic offerings at Berkeley, MIT and elsewhere must also focus on digital ethics and accountability. "If it’s going to happen anywhere, it’s going to happen in academia," she said, noting that large investments in these fields, to date, that don't address ethics are "going in the wrong direction."

The issue of ethics, she said, is being sidestepped at almost every turn by efforts like these, in favor of international economic competitiveness with China, among others. “It’s about power, it’s about control -- and it’s not at all good news for the public if we keep going like this.”

O'Neil, author of the 2016 book Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy​, noted the $1 billion price tag for the new MIT college and compared it to a much smaller, $25 million "AI for Social Good" grant program recently offered by the search and advertising giant Google.

"People who are actually worried about society and tech need a billion dollars," she said. "We need a billion-dollar investment, and $25 million isn’t big enough right now."