What has been most impactful about your time at NPS so far?

The sense of community within the Operations Research department is by far my favorite aspect of my NPS experience. I have been privileged to be a part of a large cohort that checked in together in the summer of 2024 and we have grown together and pushed each other to make the most of this amazing opportunity. The faculty has been amazing as well, they are always willing to make time for you on campus and I’ve been struck by how friendly and open they are when we run into each other in town.

Tell us about your current project Optimizing Synthetic Data Training for High Energy Lasers that use Computer Vision and the challenge it addresses?

I’ve known for a while that I wanted my thesis project to include some form of deep learning. The opportunity for this project came up and I could not have come up with a cooler application in my dreams. The technology itself is out of a movie...shooting down UAVs from space using high energy lasers. My project will increase the accuracy of classifications, key points, and aimpoints that the computer vision uses to accomplish its mission.

Your project looks at training computer vision models when data is limited. How do you see this work being applied in operational or mission settings?

Access to good data in fundamental to artificial intelligence and computer vision. In real world scenarios, our adversaries do their best to limit publicly available information about their unmanned ariel vehicles. The ability to take a limited set of surveillance photographs and build more training data through data augmentation is critical to having a robust computer vision model. The employment of Convoluted Neural Networks, Vision Transformers, and a good training set are integral to targeting UAVs from unmanned satellites.

How have the resources and relationships at NPS, including access to NVIDIA experts and infrastructure, supported your research?

The access we have to NPS high performance computing is critical to my project. The models I train would take over two weeks on my personal computer, and I am extremely grateful that I have access to NVIDIA L40S GPUs that can accomplish the same tasks in a matter of hours. I would be remiss if I did not shout out Nico Lebovitz from NVIDIA. Aside from being a good friend, he has guest lectured in a few of my academic courses, facilitated my pursuit of new certifications through the Deep Learning Institute, and has gone above and beyond to support my thesis research.

What are you hoping to achieve through the NPS project pitch session at Converge? What types of collaborations and connections are you looking for?

My pitch is the foundation of my thesis. When I saw this opportunity come up, I thought it would be a good way to push myself to make progress and get outside input. I am confident that exposure to industry partners will enlighten me and help me approach these types of problems in new ways. I have noticed throughout my career that the relationships that blossom out of events such as this have a way of persisting throughout life and you never know when you will be working with someone again down the line.