Urban computing plays a large part in Anika Tabassum’s research at the Discovery Analytics Center as she attempts to answer questions related to critical infrastructure systems: Which power grids/substations are most vulnerable and need immediate action to recover during a hurricane? Which regions are highly affected during a power outage? Are there patterns or similarities in power outages among the connected components?
Tabassum uses optimization and learning-based algorithms when trying to solve energy challenges like these. A Ph.D. student in computer science, she is also a research trainee in the National Science Foundation-sponsored UrbComp graduate certificate program, which is administered through DAC.
Critical infrastructure systems such as power, transportation, communication, and healthcare are crucial for sustaining day-to-day commodity flows vital to national security, economic stability, and public safety, said Tabassum. Failure of even a small part of such systems — caused by any natural or human-made disaster — can trigger widespread cascading failures impacting many other interdependent modules and disrupt the functionality of the entire system.
“It is challenging to understand and analyze large scale data gathered from these systems in terms of graph networks and time-series sensor technologies since they are unstructured and highly dynamic,” said Tabassum. “But extracting information like anomalies, similar patterns, and actionable insights from critical infrastructure systems can help domain experts assess, in a comprehensive manner, the complex interdependencies and failure dynamics over these systems and can also facilitate faster and less expensive decision-making,” said Tabassum.
Last summer Tabassum was a research intern at the Oak Ridge National Laboratory in Oak Ridge, Tennessee, where she applied her data mining and visualization skills in a U.S. Department of Energy (DOE) project on Smart Neighborhood. This study was accepted at the ACM International Workshop On Urban Building Energy Sensing in New York last month.
Previously, she collaborated with the Oak Ridge National Laboratory, DAC alumnus Liangzhe Chen, and her advisor B. Aditya Prakash on ‘“Urban-Net: A System to Understand and Analyze Critical emergency management.” Tabassum presented this paper in the Project Showcase at ACM SIGKDD’19, in Anchorage, Alaska, in August.
Another of their papers, “Data Mining Critical Infrastucture Systems: Models and Tools,” was published in the December 2018 issue of the IEEE Intelligent Informatics Bulletin.
With a bachelor’s degree in computer science and engineering from the Bangladesh University of Engineering and Technology, Tabassum was attracted to DAC because of the potential and interesting research in data mining and applied machine learning it offered.
“I found my advisor’s work exceptionally intriguing and very much suited to my research interest,” she said. “And once I joined DAC I found a strong collaboration of research and extremely friendly and cooperative graduate students.”
When Prakash relocated to DAC’s Arlington location in the fall, Tabassum also moved from Blacksburg.
She is on track to graduate in December 2021 and is aiming for a position as an industry researcher or an academic post-doctoral researcher.