In August, the National Science Foundation (NSF) announced the creation of a new award, meant to bring together cross-disciplinary science communities to develop foundations of data science.
Called Transdisciplinary Research in Principles of Data Science, otherwise known as TRIPODS. Phase I of this award includes a “development of small collaborative institutes,” and that is where Stony Brook University Professor Francesco Orabona comes into play.
Orabona, partnered with Lehigh University's Katya Scheinberg and Northwestern University’s Han Liu, is a member on one of 12 TRIPODS Phase I projects.
“The NSF Tripods award is a big deal for Stony Brook,” Orabona said. “This is a very big program from NSF to create collaborative institutes to create the theoretical foundations of data science.”
Orabona is the Principal Investigator on the project entitled, “Collaborative Research: TRIPODS Institute for Optimization and Learning.” This project will focus on analyzing “nonconvex machine learning models, the design of optimization algorithms for training them, and on the development of nonparametric models and associated algorithms.”
This is all part of the new NSF TRIPODS Institute, which is based at Lehigh University. The goal is to train PhD students, along with postdoctoral fellows in statistics, computer science and applied mathematics. The institute will be holding interdisciplinary workshops and winter/summer educational programs. There will also be an emphasis on deep neural networks (DNNs), with respect to specific structures of interest.
The models are transforming state-of-the-art learning tasks, such as image classification, speech recognition, machine translation and more, which are making the automated performance of such tasks efficient and reliable enough for daily use.
“Our group will focus on machine learning and optimization,” he said. “In particular, we will focus on making deep learning algorithms more efficient and to create a solid theory for them. In fact, while deep learning algorithms are everywhere in the news in these days, nobody really understand how/why they work so well.
“Isn't it scary to think that we use algorithms for critical decisions, for example self-driving cars or diagnose illnesses, and we don't understand how they work?”
The efforts are aimed at the same task: constructing efficient predictive models of high quality and high fidelity based on given data, the available sets of which may be large, complex, and heterogeneous.
“We are extremely amazed that Francesco is able to take on such a prominent role in this new program,” Samir Das, interim chair of the CS department, said. “The task at hand shows how much respect he has earned within the research community, and we are thankful the NSF is recognizing his hard work with this award.”
The NSF is one of the most prestigious organizations in the American science realm. With its dedication to supporting colleges throughout the nation, the organization continues to help advance computer science research at Stony Brook University.
Orabona is also a finalist for a Bell Labs Prize, recently being named as one of nine finalists for the award. The winner of that endowment will be named on December 8 at the Hamilton Farms Golf Club in New Jersey. He began teaching for the Department of Computer Science, which is part of the College of Engineering and Applied Sciences, in September 2016. Previously, he was a senior research scientist at Yahoo Labs and a research assistant professor at Toyota Technological Institute.
Joseph Wolkin