College of Engineering News Room

Faculty Receive NSF Grant to Detect Cyberbullying



By Brad Stager

A research team at the University of South Florida received a grant from the National Science Foundation (NSF) to explore a different approach to detecting cyberbullying in the digital communications of young adolescents.

A research team from the University of South Florida received a grant from the National Science Foundation (NSF) to evaluate the metadata of digital communications (such as time and size) instead of the content transmitted and received by adolescents.

The research team, Sriram Chellappan, PhD, from the College of Engineering and Nathan Fisk, PhD, from the College of Education, say they hope to develop useful tools such as algorithms and software applications that can predict and alert people to risky online interactions while protecting their privacy.

Currently available online behavior monitoring systems that are most widely used by parents and researchers look for keywords such as "depression," "suicide" or "hang," in content and flags them. These keyword-based systems infringe upon the privacy of users by revealing the contents of the communications. This project analyzes communication patterns of its young participants instead of message content. It is hoped this will lead to the development of new tools that respect privacy and promote early intervention when there's online abuse.

Through their research, Chellappan and Fisk will look at how the metadata associated with texts and social media postings such as the number and frequency of messages and responses, time of day and message sizes can reveal abusive online behaviors to young users and their families. The research team will recruit about 1,000 participants from among high school students in the Tampa Bay area to participate in the initial research study.

Chellappan, an associate professor in the Department of Computer Science and Engineering, said the project supports the NSF's goal of a Secure & Trustworthy Cyberspace.

"We want to understand how young people view privacy and make devices and platforms more trustworthy and secure," Chellappan said. "These are the real people who are vulnerable to this kind of thing. Our end goal is to make cyberspace more trustworthy across all age groups."

The research combines elements of Big Data due to the large volume of metadata information, and Machine Learning through the development of predictive algorithms and software applications. The work also includes elements of social science theories which deal with individual and group behavior.

Fisk, an assistant professor of Cybersecurity Education and a community and outreach liaison for the Florida Center for Cybersecurity, said separating the digital content from the research and monitoring tools developed from it helps to counter digital privacy concerns of young people and removes that conflict from the equation.

"We want to develop new ways of supervising kids in an online environment while protecting their privacy and ideally provide them with forms of intervention that they feel comfortable with," Fisk said.

One example cited is a software application that can determine when an online exchange is abusive or indicates a serious problem and can then generate an alert .

According to Fisk, the issue of promoting online safety is typically cast as requiring less privacy and that it doesn't have to be a "one or the other proposition."

"We're looking to think past a 'privacy/security' trade-off, and figure out how to give kids both at the same time," Fisk said. "Our project is intended to enroll kids to help us better understand the data we collect from them, in order to provide supervision and intervention tools that protect their privacy."

In practical terms, Fisk said the program they develop will not require parents to read what their children are writing online in order to supervise their activities. Information collected during the project will help researchers and policy makers develop tools, such as apps, that will ultimately benefit the young participants and their peers.

"The end goal of the project is to really see if we can create predictive models of abuse and then intervene for kids who have the application installed," says Fisk, who said the application he could also refer at-risk users to appropriate resources and contacts for help when abuse is detected.

Fisk also explores the issues of online safety and privacy in his book, "Framing Internet Safety: The Governance of Youth Online," published in 2016 by MIT Press.

"If there's anything I've learned it's that we were doing a really bad job of listening to kids when they were explaining what kinds of problems they have and the kinds of interventions and guidance that they needed in navigating online spaces," he said.

The $498,333 grant award for the study, "A Privacy-Preserving Meta-Data Analysis Framework for Cyber Abuse Research - Foundations," will take place over a three-year study. Learn more about the grant on the NSF website.