UC Berkeley collaborates in launching of national AI optimization institute

Photo of AI
Sujin Soman/Creative Commons
(Pixabay License: https://pixabay.com/service/license/) The National AI Institute for Advances in Optimization aims to combine artificial intelligence technology with optimization to create a more reliable and efficient energy system.

Related Posts

The National Science Foundation, or NSF, approved UC Berkeley, Georgia Tech and the University of Southern California to launch an institute with goals to fuse artificial intelligence, or AI, with optimization.

The National AI Institute for Advances in Optimization will receive $20 million over five years as one of 11 AI institutes announced July 29, according to a campus press release. The institutes join seven more that were announced last year, with funds that total $140 million in addition to the $220 million for the recently announced institutes.

“The U.S. National Science Foundation has long supported research across a variety of methods and applications in optimization,” said Georgia-Ann Klutke, the NSF program director responsible for overseeing the Georgia Tech grant, in an email. “The Institutes are the first investment at this scale for methods that explicitly integrate AI, and the first at this scale to apply optimization methods to solve complex computational problems that were previously out of reach.”

Alper Atamtürk, the UC Berkeley lead for the institute and a campus professor, said a team of campus faculty researchers in the industrial engineering and operations research department and the electrical engineering and computer science department reached out to universities with expertise in those areas to form a team.

In the summer of 2020, the team wrote the proposal, which Atamtürk said entered and succeeded in a national competition under the NSF.

Atamtürk added that the team has ambitious goals for the institute. He said the institute will use AI and machine learning to create predictions that will optimize energy generation and distribution in many areas of need such as helping airlines schedule planes or use renewable energy effectively.

“That integration with optimization and AI is going to be hugely successful,” Atamtürk said. “The institute aims to make groundbreaking research and develop some algorithms that can fuse these two technologies together for a much more resilient, reliable energy system.”

In the area of logistics systems, Atamtürk believes optimization in scheduling and planning energy use and fuel consumption can create major advances in areas of increasingly high demands such as same or next day shipping.

The institute will foster a “paradigm shift” in optimization, Pascal Van Hentenryck, institute lead principal investigator and Georgia Tech professor, said. Van Hentenryck added that the institute is building initiatives for AI education that will start in middle and high school and go up to the graduate level.

Van Hentenryck noted that the institute, through training community college faculty and creating laboratory internships, is helping minority institutions develop AI programs and bridge opportunity gaps. The institute will also expand its partnerships with historically Black and Hispanic serving high schools and colleges in various states.

Contact Andie Liu at [email protected], and follow her on Twitter at @andiemliu.