Electrical and computer engineering PhD student Michael Keyser has been awarded an NSF Graduate Research Fellowship to support his work in “sensor interfacing for analog machine learning accelerators.” This prestigious fellowship will support Keyser as he continues to research and study under ECE associate professor Armin Tajalli.
Keyser began his academic journey as a dual-majoring undergraduate at the University of Utah, studying both computer engineering and applied mathematics. After completing a Master’s degree in Electrical and Computer Engineering (in the Electrical Engineering track) at the U under Tajalli, he now pursues his doctorate under Tajalli’s mentorship.
Now an NSF-funded graduate researcher, Keyser’s first exposure to research was as a first-year undergraduate student in a biomedical engineering lab. It wasn’t until his senior undergraduate year that he began to dive deeper into computer engineering research. In the present day, Keyser’s research in analog accelerators has driven him not just to pursue a doctorate, but also to present his latest paper, “Effect of Parameter Drift on the Accuracy of Analog CNN Accelerators,” at the 2024 IEEE MWSCAS Conference where he received the Student Participation Grant.
The research Keyser presented “investigated how manufacturing and environmental conditions affect the accuracy of analog accelerators.” His work “aimed to understand how to calibrate analog accelerators for these conditions necessary to implement large-scale neural networks.”
Keyser will continue his graduate research fellowship at the U, expanding understanding of analog accelerators and their applications. We are excited to see what he does next!
Written by Marlee Jeppsen.