CIVS & ECE Host Distinguished Speaker Seminar – Mohsen Saffari

CIVS had the pleasure of co-sponsoring a Distinguished Speaker Seminar with the PNW ECE department on February 13.

Mohsen Saffari, Ph.D., Assistant Professor of Computer Engineering, presented “Smarter Grids Through Deeper Learning: Spatiotemporal Graph Models for Distribution Systems” to CIVS and PNW students, staff, and faculty. His talk was very well received and concluded with a productive Q&A session.

The increasing complexity of modern power distribution systems, driven by the large-scale integration of distributed energy resources and the need for rapid fault detection, demands advanced data-driven approaches capable of capturing intricate spatiotemporal correlations in system measurements. Traditional methods struggle with the high dimensionality of power system data, often relying on shallow architectures that fail to extract deep task-relevant features.

To address these challenges, recent research has focused on graph-based deep learning models that represent power systems as static and dynamic graphs, enabling the extraction of both spatial relationships among system components and temporal dynamics of measurements. This talk provides an overview of deep sparse spatiotemporal generative models for power system applications, with an emphasis on graph convolutional autoencoders, sparse models, and generative architectures.

Key applications, including fault classification and location in distribution systems, as well as behind-the-meter load and photovoltaic disaggregation, will be discussed along with experimental results demonstrating significant improvements over state-of-the-art methods.

About Mohsen Saffari, Ph.D.

• Professor Mohsen Saffari is an Assistant Professor in the Department of Electrical and Computer Engineering at Purdue University Northwest, where he leads the PNW-MINDs (Machine Intelligence Neural Design) Lab.
• He received his Ph.D. in Computer Science from the University of Tulsa in 2024, and his M.Sc. in Electrical Engineering from K. N. Toosi University of Technology in 2016.
• His research centers on developing trustworthy and application-driven machine learning algorithms for complex systems, with particular emphasis on spatiotemporal deep learning for power systems and intelligent transportation systems.
• Professor Saffari has published in prestigious venues including IEEE Transactions on Neural Networks and Learning Systems, IEEE Transactions on Artificial Intelligence, IEEE Transactions on Emerging Topics in Computational Intelligence, and IEEE Transactions on Industrial Informatics, with his work receiving significant recognition in the machine learning and power systems communities.