Day in and day out, faculty in the Gallogly College of Engineering at the University of Oklahoma create research opportunities for their students. Dimitris Diochnos (
photo, on left), an assistant professor in the OU School of Computer Science, has helped a student do just that.
In February, “Wind Prediction under Random Data Corruption” was showcased during the Conference on Artificial Intelligence (known as AAAI-22) sponsored by the Association for the Advancement of Artificial Intelligence. Over 40,000 researchers, practitioners, scientists and engineers took part in the virtual event Feb. 22-March 1.
The paper was co-authored by Diochnos’ former student Conner Flansburg (photo, on right) whose work was initially supported by the Research Experiences for Undergraduates program funded by the National Science Foundation.
"In this work, Mr. Flansburg showed experimentally that L1-regularization can be an effective defense mechanism for machine learning models that are used for regression and are subject to certain training-time attacks, thus complementing a similar property of L1-regularization that is known to hold in the literature regarding machine learning models that are used for classification purposes under test-time attacks," Diochnos said.
Joining OU in 2019, Diochnos serves as senior personnel for the
NSF AI Institute for Research on Trustworthy AI in Weather, Climate and Coastal Oceanography. His work examines foundational issues related to the robustness and reliability of machine learning algorithms that have applications to environmental sciences. In 2021, he published at another noted conference. “Learning Reliable Rules under Class Imbalance” was accepted at the Society for Industrial and Applied Mathematics International Conference on Data Mining.
Flansburg earned a bachelor’s degree in computer science in 2021. In February, the Broken Arrow native joined Pacific Northwest National Laboratory as a researcher in software engineering.
In 2020, OU received a $20 million grant to lead the
NSF AI Institute for Research on Trustworthy AI in Weather, Climate and Coastal Oceanography. Amy McGovern (
photo, right), an OU professor with dual appointments in the
School of Computer Science in the Gallogly College of Engineering and in the School of Meteorology in the College of Atmospheric and Geographic Sciences, is the director. At AAAI-22, she was an invited speaker for the AI Institutes' seven-member panel.