Just as machine learning has pushed forward revolutionary capabilities for facial recognition and autonomous driving, the potential for causing serious harm from tampering with algorithms poses a significant challenge. Cybersecurity and machine learning are still developing sectors, and Dr. Daniel Takabi, Associate Professor of Computer Science, is helping Georgia State students prepare to enter the increasing job market for high-demand cybersecurity skills.
Dr. Takabi has received a fellowship from Microsoft, one that will fund graduate students and post-doc opportunities within cybersecurity programs at Georgia State, which Dr. Takabi hopes will expand areas of study at the university.
It will also allow for new and innovative research projects within the department as well. The fellowship had nearly 300 applicants, and only 15 were awarded it. The research that the fellowship will fund is artificial intelligence and machine learning, Dr. Takabi said.
Dr. Takabi said that cybersecurity and machine learning are becoming more inseparable over time as information tampering and machine learning intersect.
“We can fairly easily manipulate data that would bypass a machine learning system, such as face-recognition to detect people,” Dr. Takabi said. “You can introduce some ‘noise’ to the face or image that is not visible to human eye. This would cause the machine’s algorithm to misinterpret the image.”
Some misinformation like this would be harmless. However, such an instance could also prove much more serious, Dr. Takabi said.
“This is also made evident by stop signs. There are small changes, unable to be seen by the human eye, that can cause similar ‘noise’ for self-driving cars when attempting to analyze the stop sign,” he said.
Such “noise” could disrupt the car’s ability to determine what it will do, which could lead to grave consequences. This problem is an example of the intersection of Dr. Takabi’s work, and what the Microsoft fellowship will help fund.
This fellowship will also increase the amount of activities and educational opportunities in cybersecurity at Georgia State, and Dr. Takabi hopes this fellowship with expand the horizons of scholarship in the field at Georgia State.
“We have a certificate in cybersecurity at the undergraduate level, and we also have a concentration in security and privacy at the master’s level—but we’re always trying to improve those.”
These kinds of educational advances are better preparing students, but the computer science department wanted to increase the amount of cybersecurity scholarship further. The interdisciplinary nature of cybersecurity and machine learning are reflected through the collaborations with other departments at Georgia State, such as the philosophy and ethics department. Dr. Takabi said that this is also true for the next club he helped establish, a a place for students to share and learn about cybersecurity together.
“These kinds of programs encourage students to get involved in research and academic study,” he said.
Just as machine learning and cybersecurity tend to go hand-in-hand, Dr. Takabi said he hopes this kind of educational involvement and opportunity will prepare Georgia State students for a future in the field.
Dr. Daniel Takabi is an Associate Professor at Georgia State, as well as a Next-Gen professor, and his work is centered around cybersecurity and machine learning.
—Braden Turner, Graduate Administrative Assistant, Office of the Provost