Real vs. Fake
DAIM students explore tools for managing misinformation
By David Horton '90, assistant to the dean, Artis College of Science and Technology
BEWARE FAKE NEWS! Knowing what information to trust in modern society is a challenge that only grows as time progresses. Students in the Data and Information Management (DAIM) master’s degree program at Radford University are exploring tools to help determine the level of validity of an article, video or news source.
In the 2020-2021 academic year, Michael Hammond ’21, Noah Bledsoe ’21 and Pralad Neupane ’21 explored the challenge of deepfake videos that have been edited using an algorithm to replace the person in the original video with someone else (especially a public figure) in a way that makes the video look authentic. Some fairly recent examples include actors posing as celebrities like Tom Cruise or Arnold Schwarzenegger — the actors’ faces are overdubbed with those of the celebrities, then shared across the internet.
Through several different projects, the students researched innovative tools such as edge enhancement, convolutional neural networks and other detection procedures that can help give hints as to the validity of the video in question.
This fall, Amanda Tolman ’21, a graduate student, is conducting a capstone project that will look for patterns within various social media posts to predict if false content is present. A trustworthiness score will also be calculated to help users decide whether to trust a source or the information shared. With further research, this score could alert readers to potentially false information.
Through this technology, DAIM students hope to develop procedures to assist consumers in determining which items are trustworthy and which raise concerns about validity.
“Data engineering builds solutions to many emerging problems, like identifying misinformation and security vulnerabilities, and helps people and organizations make better decisions,” said Jeff Pittges, Ph.D., DAIM program coordinator and interim director of the School of Computing and Information Sciences.