Eying His Future and Teaching Machines to See
USF graduate student Lucas Miller is trying to solve a problem that even the most advanced AI models have yet to crack: How to help machines understand and interpret video. A software engineer at AssistRx, Miller is in his final semester of the Master of Science in Computer Engineering program. Like many adult learners, he balances full-time work with a rigorous academic schedule. What sets him apart is his focus on thesis research that pushes the boundaries of what today鈥檚 artificial intelligence can do.
A Research Question That Matters
Miller鈥檚 thesis explores how artificial intelligence can be trained to recognize and interpret sequences of human actions in video. While today鈥檚 language models can summarize text and describe images, they struggle to make sense of moving scenes over time. His work in video segmentation is aimed at helping machines analyze what people are doing, not just what objects appear in a frame.
鈥淚t is a huge step from understanding a still image to understanding video,鈥� he said. 鈥淐urrent models, they can鈥檛 understand videos. So, my thesis project work, we're trying to get the computer to understand sequences of events, viewing them as human actions, not just objects.鈥�
Using the surveillance and security industry as an example of how it might be applied in industry, Miller said that it could be a tool for workers. It isn鈥檛 about the computer watching people, he said, but about helping reviewers know where to look more closely. 鈥淔or example, if a tech is reviewing footage, it can say,鈥� hey, maybe this part you should look at closer,鈥� and tell you what鈥檚 happening.鈥�
Miller said the work could have far reaching implications and could be applied to many industries, from healthcare to robotics, education, or entertainment.
鈥淭his isn鈥檛 just theory,鈥� he said. 鈥淭hese are the kinds of problems that tech companies are actively working on.鈥�
Rigor and Support in the Master鈥檚 Program
Miller said he always knew that he would work in the computing field. He earned an undergraduate degree from Florida Polytechnic University and came to USF to deepen his skills and explore AI research at a higher level.
鈥淚鈥檝e been programming computers since I was twelve years old, and every day I try to push the boundaries of what I know and understand,鈥� he said.
While he admits balancing full-time work and graduate study requires discipline, he says the payoff is worth it.
鈥淭he program is demanding, but that鈥檚 what makes it effective,鈥� he said. 鈥淭he core courses give you the foundation, and then the thesis work helps you apply it in ways that are creative, ambitious, and sometimes unpredictable.鈥�
He credits much of his growth to mentorship from faculty like Distinguished University Professor Sudeep Sarkar, a renowned expert in computer vision. Working under Sarkar鈥檚 guidance, Miller said he gained both technical insight and confidence as a researcher.
鈥淗e challenges me to think deeper and sharper, but he also gives me the freedom to explore,鈥� Miller said. 鈥淭hat balance has been incredibly valuable.鈥�
Miller uses words like engaging, relevant, interesting, and dynamic to describe the classes. He appreciates the structure the energy faculty bring to the classroom.
鈥淭he professors do a good job keeping everyone interested. Even the classes that are more theory-heavy feel engaging,鈥� he said.
They know how to take a complex concept and connect it to something that鈥檚 real or useful, he added, and do so in an interactive, discussion-based space.
鈥淥ne professor even joked, 鈥楲ucas, stop answering questions鈥攍et someone else try,鈥欌�� he said. 鈥淭hey keep things light but make sure everyone鈥檚 engaged.鈥�
Looking Ahead
As he works to teach computers how to see and interpret the world, he says the experience at USF has helped him see his own path with new clarity. He plans to pursue a PhD next and then hopes to land a research-focused role in the tech industry.
In April, he was awarded USF鈥檚 Graduate Student Success Fellowship, which provides tuition support and a stipend for up to three years of doctoral study. He also received a college-sponsored award, the University Graduate Fellowship, which supports incoming doctoral students. The awards and recognition, he says, reinforced his belief that he's on the right track.
鈥淚 have to get this thesis written first,鈥� he joked. 鈥淏ut I am really excited about what鈥檚 next. It isn鈥檛 easy, but this program helped me grow a lot,鈥� he said. 鈥淚t showed me that I can push myself, and that I can do meaningful work.鈥�
With graduation in sight and support for the next chapter already in place, Miller is ready to continue the work that inspired him from the start: Teaching machines to see while keeping his eyes on his own future.