PhD student knows how to (neural) network

People Mentioned in This Story

Maryam Parsa

Body

Electrical and Computer Engineering PhD student Shay Snyder is not only "a very capable student," as his advisor assistant professor Maryam Parsa said, but he is also a tenacious networker. Passionate for his subject and tireless in his pursuit of research opportunities, Snyder joined George Mason University's Department of Electrical and Computer Engineering (ECE) and assisted in securing at least two years of funding for a project in Parsa's lab.

Parsa's research focuses on neuromorphic, or "brain-inspired," computing across various levels of the computing stack, Parsa explained. The question is, 'How can neuroscience and biology help us redesign computers to be more capable, energy-efficient, and robust?,' she said.

Although Snyder began his PhD in January 2023, he had been part of the Mason Nation for the prior year, working as an undergraduate research assistant for Parsa as he finished up a bachelor's degree at East Tennessee State University.

Snyder was an effective networker even as a teenager. He attended a two-week summer camp at Oak Ridge National Laboratory as a junior in high school and was eager to return.

Shay Snyder. Photo provided by Maryam Parsa.

"I found out that they have an internship program for high schoolers, so my entire senior year of high school, I emailed and emailed, saying like, 'Hey, I like this about your work, and I want to work with you.'" He recalled, "I sent so many emails, I got two e-mail accounts blocked because they thought I was spamming them, but luckily I heard back … so a week after I graduated from high school, I got to go to work at Oak Ridge for a summer, and that kind of started everything." He added, "It was a miracle."

The internship was an eye-opening experience for Snyder. He explained, "Where I grew up in East Tennessee was much more remote. There are not a lot of technology jobs, and coming from my high school, I think over 50 percent of the kids were below the poverty line. So getting to see that was a whole new world for me, and I was determined that I wanted to continue doing work like they do at Oak Ridge."

After that initial internship, Snyder continued reaching out to scientists he had met or learned about at Oak Ridge. He attended Lunch and Learn events for undergraduate students, at one of which he met Catherine Schuman, who introduced Snyder to the concept of neuromorphic computing.

"I was super inspired," he said. He followed up with Schuman and was offered a yearlong position at Oak Ridge doing autonomous vehicle research for General Motors. He took his undergraduate classes part-time and worked remotely to leverage this opportunity.

Snyder first encountered Maryam Parsa when she gave a presentation at the International Conference on Neuromorphic Systems. She was then a post-doc also working at Oak Ridge National Laboratory with Schuman. Snyder contacted her, and the two stayed in touch. Eventually, after joining the ECE faculty, Parsa hired Snyder remotely as an undergraduate research assistant while he finished his undergraduate degree. (Meanwhile, Snyder also started a part-time internship at the Pacific Northwest National Laboratory.)

"[Snyder] was still an undergraduate, and he worked on an area that was completely different from his background," said Parsa, emphasizing, "The project was on neuromorphic computing, but it involved learning multiple new areas including Bayesian learning and Physics-Informed Neural Networks (PINN). He quickly learned the necessary background and developed Lava Bayesian Optimization that later merged with the Intel's neuromorphic optimization framework, Lava-Optimization." The work has been published in multiple conference papers at the International Conference on Neuromorphic Computing Systems (ICONS), the Great Lakes Symposium on Very Large Scale Integration (GLSVLSI), and the International Conference on Machine Learning and Applications (ICMLA), and Snyder's contribution is publicly available in Intel's open-source Lava framework.

Two years ago, after Parsa circulated a call for proposals from the Army Ground Vehicle Research Center with her lab, Snyder approached her with an idea involving hyperdimensional computing. She encouraged him to write a draft. The two refined the idea and submitted a white paper on the topic. They were then invited to submit a full proposal, which was accepted for funding. So, the same semester that Snyder joined Parsa's lab as a PhD student, the lab also began to receive funding for a project based on Snyder's idea.

Maryam Parsa and lab members. Photo provided by Parsa.

Since Spring 2023, Snyder has become a full PhD student at George Mason and remains an important piece of Parsa's lab. Snyder has taken lead on multiple projects, academic, industry and government collaborations, and published multiple papers. He has also been mentoring not only high school interns and undergraduate students in Parsa's lab, but also other PhD students. Since Spring 2022 when Snyder joined Parsa's lab, he has been the first author of five published conference papers in top artificial intelligence and neuromorphic venues and co-author in one published journal paper and six published conference papers.

Parsa deeply appreciates having Snyder in her lab, not only for his technical and professional skills but also for the positive energy he brings to the team. She highlights how Snyder's can-do attitude, curiosity, and collaborative nature have been instrumental to the success of numerous projects. His peers frequently rely on him, valuing his contributions and leadership in driving projects forward. Parsa regards working with Snyder as a true pleasure, given the impact he has made inside and outside in the lab.