Professor Li created ImageNet, a large dataset that would change the machine learning world forever.
When her colleagues were working to create models to help computers detect and classify images, Li found that they had to create a separate model for identifying cats and another for dogs, for example. She wondered whether a large dataset of all the images in the world can be created. At first, her project attracted little attention, but later on, she was able to gather enough help to assemble over 15 million images. She then used her dataset as part of the ImageNet Large Scale Visual Recognition Challenge. Before ImageNet, computer vision accuracy rates were no more than 85%- afterwards, many models reached over 97% accuracy. ImageNet is so powerful that it is used in self-driving cars, facial recognition, phone cameras, and more.
Li turned AI, a seemingly isolated field, into an interdisciplinary one. She partered with a medical school to use AI in a camera system to monitor the hand-washing stations to make sure all hospital workers wash their hands properly. A doctor at the hospital says Li was the only one from the computer science department to reach out to him.
Because Professor Li is frequently the only women of color among colleagues and peers, she started a nonprofit organization to increase diversity and inclusion in the AI industry and encourage underrepresented groups to learn more about the field of AI.
AI4ALL became a summer camp offered at Stanford, Carnegie Mellon, UC Berkeley, and other college campuses for underrepresented and marginalized groups such as girls, people of color, and people from disadvantaged backgrounds who are in high school. Over 50% of AI4ALL's summer program participants are female and over 90% are students of color. ✦ Read more about AI4ALL here. ✦