John Wright, Principal Investigator in NSF Funded TRIPODS Institute

September 11, 2017
Prof. John Wright

Data science is in its adolescence – a field whose adult foundations have yet to be established – akin to the state of computer science in the 1960s. And now, thanks to the National Science Foundation (NSF), the Data Science Institute (DSI) will be the place to cement the foundations of the emerging field.

The foundation announced today that it will fund the Transdisciplinary Research in Principles of Data Science (TRIPODS), an institute that will be based at DSI. John Wright, an associate professor of Electrical Engineering at Columbia, who is the principal investigator on the NSF grant, says TRIPODS will focus on three areas: research, teaching and center-building. The TRIPODS Institute will be a mecca of data science – a hub where statisticians, mathematics, computer scientists and others will gather to establish the theoretical foundations of the field.

“Data is accelerating the pace of scientific discovery and innovation,” says Jim Kurose, NSF’s assistant director for Computer and Information Science and Engineering, “and TRIPODS will enable continued data-driven discovery and breakthroughs across all fields of science and engineering.”

Technological advances have allowed for a deluge of data. And the availability of these data – their volume and the speed at which they are amassed – is transforming science and engineering. By harnessing this data revolution, one of the “10 Big Ideas for Future NSF Investments,” the foundation seeks to support data-driven research while developing the next generation of data scientists.

The TRIPODS award will fund major investments at the Data Science Institute. Those investments include state-of-the-art mathematical and statistical tools, better data mining and machine learning approaches as well as enhanced data-visualization capabilities.

The research initiatives include gaining a deeper understanding of optimization problems, designing computational data models, and developing interactive protocols in data science. The educational aspect includes developing data science curricula for master’s and doctoral programs, with a focus on interdisciplinary courses to build a common language for a new generation of data scientists and engineers. And center-building activities will include workshops, summer schools, lecture series, scholarly visits and outreach activities---all of which will articulate a universal language for foundational research and education in data science.

“It’s a fascinating and exciting time to be working on data science,” says Wright, a DSI professor. “Data science calls upon a number of disciplines but is a distinct discipline. Research in the field holds so much promise. Those of us at the Data Science Institute have such fantastic motivation -- we have such cool research going on here -- it’s wild. And TRIPODS will be the international center for the exchange of ideas.”

Alexandr Andoni, another professor at the Data Science Institute, said he’s excited about the goal of TRIPODS, which is to establish the unified foundations of data science. Data science is a nascent area – a new language where many disciplines can say something, with computer science, statistics and mathematics having particularly much to say – each contributing its own perspective of data science problems, he says.

“It’s particularly surprising to discover how different disciplines describe the same phenomena in completely different languages,” adds Andoni, a professor of Computer Science at Columbia Engineering. “Where are these perspectives really the same ones, where are they different, where are they complementing each other, and where can one perspective be further enriched by fusing elements from the others? Bridging these different languages and perspectives is precisely the goal of TRIPODS.”

Andoni, an expert on the algorithmic foundations of massive data sets, used an exploration metaphor to explain the palpable sense of excitement and novelty surrounding the pioneering field of data science.

“To use another parallel for the goal of developing the foundations of data science, it’s bit like settling a new planet where you have to send a team with a multitude of complementary skills to be successful,” he says. “Hence I see TRIPODS as an excellent platform to bring these new disciplines together, learn from each other, and ultimately team up to develop the nascent area of data science. I’m very excited to be part of it.”

-By Robert Florida

Original article can be found here.