You are here
Ioannidis Awarded NSF CAREER Grant
ECE Assistant Professor Stratis Ioannidis was awarded a $460K NSF CAREER grant for "Leveraging Sparsity in Massively Distributed Optimization".
The unprecedented growth of the Internet and social media have created enormous volumes of data—about consumers, their browsing habits, their shopping preferences, and other online behaviors. This information holds significant strategic and financial value for manufacturers, retailers, and other businesses. But the question is, how can this information be effectively mined for valuable insights? How can the critical data be distinguished from the trivial?
According to Stratis Ioannidis, an assistant professor in the Department of Electrical and Computer Engineering (ECE) at Northeastern, data management experts have two options. “We can apply traditional methods of data analysis and computer processing, and wait months or even years to arrive at meaningful answers,” he explains. “Or we can develop more innovative methods of managing Big Data—including new processing schemes and mathematical algorithms—that accelerate results.”
Since he joined the College of Engineering in 2015, Ioannidis has been leading a team of researchers focused on Big Data innovations. In his Data, Networks, and Algorithms Lab, he is working to automate and accelerate the process of sifting through billions of data points to discover those pieces of information that truly matter.
Recently Ioannidis won a prestigious CAREER Award from the National Science Foundation (NSF) to fund this research. “One of the key concepts in managing Big Data is parallel processing,” Ioannidis notes. “By spreading complex calculations across thousands of computer processing units, or CPUs, we can rapidly identify trends, patterns, and critical insights in even the largest databases. The problem is that many mathematical algorithms don’t lend themselves to parallel implementations across multiple computers.”
To address this challenge, Ioannidis is investigating new algorithms and computing platforms that enable parallel solutions for common Big Data problems. He and his team are leveraging the Massachusetts Green High Performance Computing Center (MGHPCC) to test new algorithms and processing methods across hundreds of machines and thousands of CPUs. The MGHPCC is a joint venture between Northeastern and four other universities—Boston University, Harvard, MIT, and the University of Massachusetts.
While understanding consumers’ online behaviors and preferences is one application of this research, Ioannidis emphasizes that his work has broad implications for industry, academia, and government agencies. “Whenever there is a huge volume of information that must be sorted and mined, new algorithms and processing schemes can add tremendous value,” he says.
Ioannidis, who left Yahoo Labs to join Northeastern, has always enjoyed problem solving—and he believes Northeastern is the perfect environment for his research on massive parallelization.
“From resources like the MPGHPCC to the interdisciplinary collaborators that I can call upon across the university, Northeastern is an ideal place to tackle big, sophisticated challenges,” Ioannidis states. “Solving multifaceted problems requires a varied skill set and a diversity of perspectives. Northeastern has assembled all the leading expertise I need to support my research, and I don’t even have to leave the campus.”
View the abstract from the NSF