1/13/2023 Aaron Seidlitz, Illinois CS
Illinois CS professor Derek Hoiem, nominated by fellow professor David Forsyth, earned this acknowledgment in December, calling it his most significant recognition in the field thus far.
Written by Aaron Seidlitz, Illinois CS
Since 2009, Illinois Computer Science professor Derek Hoiem has consistently delivered relevant research and entrepreneurial energy to the study of computer vision, a subset of Artificial Intelligence.
His success in this area has led to several career recognitions, including the Sloan Research Fellowship, IEEE PAMI Young Researcher Award, NSF CAREER Award, University Scholar, and more.
Then, in December, Hoiem was further honored for his dedicated and impressive research, as he was elevated to IEEE Fellow for his contributions to computer vision. The IEEE is, of course, a leading community dedicated to “advancing technology for humanity.”
Of his career acknowledgments thus far, Hoiem considers this elevation to IEEE Fellow “the most widely recognized as a significant achievement.”
Outside of the prominence of the organization, he feels this way because of the way IEEE selects its fellows based on quantifiable evidence of impact. In that sense, this recognition has been a long time in the making, as Hoiem’s interest in computer vision and its potential impacts on our society stems from his time as undergraduate student with the University at Buffalo – the State University of New York (SUNY).
“I’ve long been fascinated by intelligence and the idea artificial intelligence,” Hoiem said. “At SUNY Buffalo, I was the only undergraduate in the machine learning class, which wasn’t nearly as popular a topic then. As an independent study, I learned some basic image filtering operations to count Moiré fringe patterns to measure stress on a computer chip.
“Then, I became interested in computer vision as one of the more compelling applications of machine learning, and I’ve been hooked ever since.”
His genuine interest has held strong for years now due to the impact vision has on everyday aspects of life.
“Computer vision, and related areas, is already used in daily life through camera phones, vehicle safety, biometrics, and other areas,” Hoiem said. “Vision is one necessary component to human-like intelligence, and has been called AI-complete, meaning that solving computer vision requires solving AI. Indeed, we are finding that advances in computer vision, natural language processing, and other areas are often mutually beneficial, so the barriers between different subfields of AI are breaking down.”
Here at Illinois CS, Hoiem credited fellow Illinois CS professor David Forsyth – who nominated him for elevation to IEEE Fellow – with playing a major role in his own development.
Their often collaborative efforts have pushed the computer vision group within this department forward, a notion that Hoiem said is due to Forsyth’s leadership.
“David has taught me a lot about how to teach and how to mentor, and I’ve also learned a lot through our research collaborations,” Hoiem said. “Besides possessing one of the most fantastic minds, David is exceptional for his kindness and generosity, and it’s a true privilege to be his colleague and have his support.”
The reach of Hoiem’s impact in computer vision has delivered outside of academia, too.
This has largely come through Reconstruct, a company he co-founded with Mani Golparvar Fard, a professor with Civil & Environmental Engineering, and Tim Bretl, a professor with Aerospace Engineering.
The company’s growth showcases the relevance of computer vision in the construction industry and has been effective enough to earn more than $17 million in Series B funding a year and a half ago.
“Construction is a trillion-dollar industry in the United States. Thirty percent of the costs are due to failures in coordination. There have been very minimal improvements in productivity over the last 50 years, while other sectors like manufacturing improved quite dramatically. The construction industry was ready for new solutions,” Hoiem said in an article from The Grainger College of Engineering’s magazine Limitless.
Moving forward, there are a number of ways Hoiem wants to further investigate vision and its impacts.
“Currently, I’m most excited about general purpose learning – i.e., creating models that can do many tasks that they weren’t specifically designed to do – and about the potential for neural rendering fields to create 3D semantic models that bring AI to the physical world like never before,” Hoiem said.
Read more IEEE Fellow coverage from The Grainger College of Engineering.