Fei-Fei Li
Why is Fei-Fei Li known as the “Godmother of AI”?
What is ImageNet and why is it significant?
What is AI4ALL and what is its mission?
What are some of Fei-Fei Li’s notable awards and honors?
Fei-Fei Li (born July 3, 1976, Beijing, China) is a Chinese American computer scientist widely known as the “Godmother of AI” for her groundbreaking work in computer vision, which serves as a foundation for many image-recognition artificial intelligence (AI) systems.
Early life and education
Li was born to Kuang Ying, a secondary-school teacher, and Li Sun, a computer department employee at a chemical factory. She was raised in Chengdu, in Sichuan. As a child Li excelled at school and outperformed the boys in her classes, which teachers viewed as improper. When Li was 12, her father emigrated to the U.S. To ensure that Li had access to better educational opportunities and to escape China’s political discord, she and her mother emigrated to join Li’s father in 1992 when Li was 16. The family lived in a one-bedroom apartment in Parsippany, New Jersey, and rebuilt their lives.
Li’s father worked as a camera repairman while her mother was a supermarket cashier. Li attended Parsippany High School, working part-time as a waitress and cleaner to help her parents. She credits her high school math teacher Bob Sabella with helping her adapt to a new language; he and his wife also helped Li’s family by lending her parents money to purchase a dry-cleaning business. In 1995 Li was awarded a scholarship to Princeton University. She attended classes during the week and traveled home on the weekends to help her parents run the business. She earned a bachelor’s degree in physics in 1999.
Li began her doctoral studies in 2000 at the California Institute of Technology, where she worked on projects in computer science, electrical engineering, and cognitive neuroscience. In 2005 she received her Ph.D. in electrical engineering. During her studies, Li played a pivotal role in advancing “one-shot learning,” a method applied in computer vision and natural language processing that enables AI to generate predictions with limited data.
Academia and ImageNet
Li began her teaching career in 2005 at the University of Illinois at Urbana-Champaign, where she was a professor of computer and electrical engineering and a full-time faculty member of the AI working groups at the university’s Beckman Institute.
In 2006 Li realized that the rapid growth of online image data could offer opportunities to develop more advanced and reliable models and algorithms for organizing, indexing, retrieving, and interacting with visual and multimedia data. In 2007 Li joined Princeton University’s computer science faculty.
At Princeton, Li began creating ImageNet, collaborating with other computer scientists, including Princeton professors Jia Deng, Kai Li, and Wei Dong. By 2009 the team believed that its collected and annotated database of 3.2 million images was comprehensive enough to help advance AI applications, and they published the paper “ImageNet: A Large-Scale Hierarchical Image Database,” along with the database.
The following year, Li and her colleagues held a competition in which they encouraged other researchers to accurately classify as many of the ImageNet images as possible, with the lowest error rate, using their own algorithms. Each year, the error rate of the researchers’ algorithms decreased, improving because of their training on the ImageNet dataset. ImageNet grew to contain nearly 15 million images, with close to 22,000 indexed synonym sets, which are the groupings used to describe each individual concept. Considered crucial to the development of AI image-recognition systems, ImageNet is often referred to as the “eyes of AI.”
In 2009 Li became an assistant professor at Stanford University. In 2013 she was named director of Stanford’s AI laboratory, and she was promoted to full professor in 2018. During a sabbatical from Stanford, she worked as a vice president at Google and chief AI/ML (machine learning) scientist for Google’s Cloud division.
“I am a shy person, not good at expressing myself, but I insisted on publishing a book because the field of AI cannot lack female voices.”
—Li speaking about her memoir to the South China Morning Post
Li returned to Stanford in 2019 and was appointed co-director of the new Stanford Institute for Human-Centered Artificial Intelligence (HAI). To support HAI’s mission of ensuring that humanity benefits from AI, Li researched how ambient intelligence technology, or responsive technologies that use sensors to analyze environments, could be used to improve healthcare systems. Ambient intelligence systems can improve patient care and provide physicians with real-time access to medical records, lab results, and imaging studies.
- Kusaka Memorial Prize in Physics, Princeton University, 1999
- Fellowship for New Americans, Paul and Daisy Soros, 1999–2003
- Distinguished Alumni Award, California Institute of Technology, 2020
- Time AI100, Time magazine, 2023
- Woodrow Wilson Award, Princeton University, 2024
Li is a vocal proponent of inclusion and representation for women and minorities in computer science. In 2017 she cofounded the nonprofit organization AI4ALL, which calls for inclusivity in AI education. AI4ALL offers Stanford’s AI courses to high school students to encourage young women and minorities to consider computer science-related careers. Li also speaks about the importance of equality, diversity, and inclusivity in the technology industry in her keynote speeches at academic and industry events. In November 2023 she published a memoir, The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI.