October 12, 2022
UW’s Yejin Choi wins MacArthur Foundation ‘genius grant’
Yejin Choi, University of Washington professor in the Paul G. Allen School of Computer Science & Engineering, has been named one of this year’s MacArthur Fellows. Choi uses natural language processing to develop artificial intelligence systems that have the ability to reason and can understand the implied meanings in human language.
Yejin Choi, University of Washington professor in the Paul G. Allen School of Computer Science & Engineering, has been named one of this year’s MacArthur Fellows.
The fellowship from the John D. and Catherine T. MacArthur Foundation comes with an $800,000 stipend, commonly known as the “genius grant,” for recipients to use as they see fit. The Chicago-based foundation announced the 25 fellows on Wednesday.
Choi uses natural language processing to develop artificial intelligence systems that have the ability to reason and can understand the implied meanings in human language. AI often uses rules-based models, such as logic or probability. But Choi says that these rules are too rigid to make sense of the nuances that most people take for granted when they talk to each other.
“When I received the phone call from the Foundation, I thought they were going to ask me to do some consulting work,” Choi said. “My heart almost stopped beating when I heard ‘congratulations’ instead. This is such a great honor because there have been only two other researchers in the natural language processing field who have received this award.”
Choi is still working on the exact plans for the award, but hopes to use it to pursue impactful, though potentially risky, research ideas.
“Taking the road less traveled may seem exciting at first, but sustaining this path can be lonely, riddled with numerous roadblocks and disheartening at times,” Choi said. “This fellowship will power me up to go ahead and take that adventurous route.”
Choi has already made advancements in several areas to push the field of natural language processing forward. One example is combining both visual and text inputs for these systems. Traditionally, these models are trained solely with text inputs, but Choi has designed models with both text and image inputs that reinforce each other to better mimic how people acquire knowledge about the world.
In another line of work, Choi uses computational linguistics to help AI identify deceptive intent or sentiment in writing. For this project, the research team designed a method to automate accurate detection of fake online consumer reviews. Then Choi extended this work to include assessing news articles based on intent to deceive as well as categorizing the articles as “hoax,” “satire” or “trustworthy.”
Recently Choi’s team developed Ask Delphi, a research prototype designed to make AI more ethically informed. When presented with a moral dilemma — such as ignoring a supervisor’s phone call during working hours — Delphi weighs in on whether the situation is OK. Choi led the Delphi project through a joint appointment at the Allen Institute for Artificial Intelligence.
In announcing the award, the MacArthur Foundation said, “Choi’s research brings us closer to computers and artificial intelligence systems that can grasp more fully the complexities of language and communicate accurately with humans.”
After receiving a doctoral degree from Cornell University, Choi was an assistant professor in the computer science department at SUNY Stony Brook before joining the Allen School faculty in 2014.
“The MacArthur Foundation could not have picked a better candidate than Yejin Choi,” said Magdalena Balazinska, UW professor and director of the Allen School. “Yejin epitomizes what the ‘genius grant’ is all about — she is fearless about breaking down barriers, asking hard questions and pushing AI in exciting new directions.”
“Natural language processing, and AI more broadly, have become deeply intertwined with all aspects of society,” Balazinska continued. “It is critical that we deeply understand its capabilities and limitations, and that we push those capabilities in the interest of social good. Yejin does exactly that, and she does so in such creative and sometimes astonishing ways. I can’t wait to see what she will accomplish next, thanks to this award.”
In winning the MacArthur Fellowship, Choi joins 13 others who were current University of Washington faculty at the time of their awards. The most recent winner was Trevor Bedford, affiliate associate professor of genome sciences and of epidemiology at the UW, who received the fellowship in 2021.
“We are delighted and proud that Yejin has received this significant acknowledgment of her innovative leadership in AI research,” said Nancy Allbritton, Frank & Julie Jungers Dean of the UW College of Engineering. “Yejin is a creative, gifted computer scientist whose expertise in natural language processing is advancing AI systems’ capacities. We are eager to see how her work will further transform computer-human interactions and advance engineering for the public good.”
The fellowship is awarded, in the words of the Foundation, to “talented individuals in a variety of fields who have shown exceptional originality in and dedication to their creative pursuits.” Winners have been nominated anonymously by leaders in their fields and chosen by an anonymous selection committee.
“It’s been several weeks since I learned about this award, and it still feels so surreal,” Choi said. “I feel like the universe is sending me this message that it’s OK to be imperfect, it’s OK to try many ideas and fail along the way, as long as I continue learning and don’t give up.”
For more information, contact Choi at yejin@cs.washington.edu.
Tag(s): awards • College of Engineering • Paul G. Allen School of Computer Science & Engineering • Yejin Choi
