🖼️

Fei-Fei Li

Original nameFei-Fei Li

American Computer Scientist, Creator of the ImageNet Dataset, "Godmother of AI"

The Names Remembered · Small Roles, Famous Names

The ImageNet DatasetThe ImageNet Large Scale Visual Recognition ChallengeAdvocacy of Human-Centered AI

Who they are

Fei-Fei Li (1976– ) is an American computer scientist born in Beijing, China, who emigrated to the US with her family in her youth and thereafter long taught at Stanford University. She keenly realized that computer-vision research of the time was widely limited by too-small and too-uniform training data, and from 2007 led a team to build an unprecedented large-scale image dataset "ImageNet," over several years and with large-scale crowdsourced annotation at last building a vast dataset of over 14 million images systematically annotated across more than twenty thousand categories. She thereafter from 2010 launched the "ImageNet Large Scale Visual Recognition Challenge," inviting research teams worldwide to compete publicly on the accuracy of image-recognition algorithms on this unified dataset. It was precisely on this competition platform that in 2012 the "AlexNet" designed by Hinton’s team won overwhelmingly, directly triggering the full rise of deep learning in computer vision thereafter — without the large-scale annotated data ImageNet provided as "fuel," this decisive breakthrough of deep neural networks might well have taken longer to come. Li thereafter long advocated the idea of "human-centered AI," stressing that technological development should take advancing human welfare as its core goal, and co-founded the AI-education non-profit "AI4ALL" oriented to disadvantaged groups, and for these contributions is often called by the media the "godmother of AI."

Primary sourcesDeng et al., "ImageNet: A Large-Scale Hierarchical Image Database" (2009 paper)

Key stories

A "Textbook" Prepared for Machine Learning

Li realized that a human infant learns to recognize the things of the world only through tens of thousands of observations in daily life, and that a computer-vision algorithm might likewise need to encounter a large enough, diverse enough body of image data to truly "learn to see," and she therefore launched the building of ImageNet, an unprecedented large-scale image dataset, over several years organizing tens of thousands of annotators worldwide through the Amazon Mechanical Turk crowdsourcing platform, at last building the "data fuel" thereafter proven indispensable to deep learning’s visual breakthrough.

A Competition That Witnessed the Start of a New Era

The ImageNet competition Li launched was held yearly from 2010, drawing top research teams worldwide to compete publicly, and in 2012 the AlexNet designed by Hinton’s graduate team achieved in this competition an accuracy far exceeding all previous entrants, a dramatic victory thereafter acknowledged as the landmark moment the deep-learning era truly arrived — and all of it built on the foundational work of Li quietly building the dataset in the years before.

Relationships

Disciple

🧠Geoffrey Hinton　AlexNet proved the overwhelming advantage of deep learning precisely on the ImageNet dataset and competition platform Li founded

Echoes today

Below are how modern works borrow or reinterpret this name or story — not the original material. The two differ, so keep them apart.

The "godmother of AI" title and "human-centered" advocacyLi has in recent years continued to advocate the idea of "human-centered AI," and published the memoir The Worlds I See recounting her study and research experience, one of the most publicly influential women scholars in AI today.

Appears in

The Explosion of Deep Learning and Generative AIPerseverance in the "Neural Network Winter" and the Shock of AlexNet

← Back to Figures