College of Engineering News Room

Computer Science and Sociology Researchers Team Up for DARPA Grant

The group will study how fast information travels in online environments.

by Brad Stager

News travels fast on the Internet and a $1.7 million grant from the Defense Advanced Research Projects Agency (DARPA) will fund researchers from USF's College of Engineering and Department of Sociology in the College of Arts and Sciences to examine how fast and in what ways different kinds of information travel through online environments, including social media.

The transdisciplinary project, called "Modeling Information Diffusion Processes with Deep Learning Algorithms," is part of the government agency's Computational Simulation of Online Social Behavior (SocialSim) program, which according to DARPA's website, seeks to accurately simulate online social behavior through computer modeling. For more than 50 years DARPA, an agency of the U.S. Department of Defense, has held to a singular and enduring mission: to make pivotal investments in breakthrough technologies for national security.

In announcing the grant award, Professor and Chair of USF's Computer Science and Engineering Department Sudeep Sarkar summarized the project's purpose as evaluating "deep learning methodologies using neural networks for predicting how information spreads in various online environments." Sarkar says working on a DARPA project is a big deal.
"We're very excited. DARPA grants are very hard to get. DARPA picks projects that are at the very cutting edge."
Sarkar adds the program is about all the factors that may (or may not) govern the spread of information across a variety of online platforms. Potentially, billions of users can be applied in areas such as public health and emergency management. He characterizes the project as "STEM research involving the humanities."

"Two disciplines are coming together from their comfort zones to solve a real social problem." Sarkar adds that this kind of cooperation between academic and scientific disciplines is a research trend that he expects will continue to grow.
Representing those two disciplines are Computer Science and Engineering Professor Anda Iamnitchi, who is the primary investigator for the grant and her collaborators, Distinguished University Professor of Computer Science and Engineering Lawrence Hall and Distinguished University Professor of Sociology John Skvoretz.

The use of large-scale distributed computing systems to explore online social networks and related phenomena is Iamnitchi's research focus at USF and reflects her interest in social science as well as technology. Iamnitchi cites a 2016 survey by the Pew Research Center (conducted in association with the John S. and James L. Knight Foundation) showing 62 percent of adults in the United States receive news via social media channels as evidence of the need to understand how information travels in online environments. She says it's a new field of research with many challenges.

"There's been a lot of work in this area but not a lot of validation."  Iamnitchi says her team will employ deep-learning neural networks to model complex nonlinear relationships such as those found in online environments.

The grant's co-investigators each bring particular skills and perspectives from the computer and sociology fields to do the work. Iamnitchi has collaborated with Skvoretz on many research projects, including a study on the propagation of cheating in online gaming, and values what sociology can bring to STEM research. "This is a much older science than computer science," says Iamnitchi. "Sociologists understand the phenomena of social networks."

Skvoretz works the quantified end of his field as a mathematical sociologist with research interests in areas such as social psychology, as well as network analysis and modeling.

"It's an interesting challenge that stretches your thinking about problems and that's a good thing," says Skvoretz.
Developing algorithms that illustrate and predict online data flow through social networks requires "connecting" a lot of dots, or neural network nodes, and Hall's expertise in machine learning, data mining and pattern recognition will help the research team accomplish that. Hall says expanding the knowledge of how information flows through the online environment can provide benefits to society and the individual person.

"It could help you know a little bit better what you should pay attention to, which is more important, and potentially more impactful."

The DARPA grant will keep Iamnitchi and her team busy researching social media data flow for the next four years. She says it's the kind of project that keeps her job interesting.

"It's exciting because it's not easy."