How did a Stanford professor's ambitious vision to catalog every object in the world transform AI?

The answer

In 2009, Stanford professor Fei-Fei Li and her team released ImageNet, a massive visual database containing over 14 million hand-annotated images organized into more than 20,000 categories based on the WordNet hierarchy. Li used Amazon Mechanical Turk to crowdsource the labeling effort, with workers annotating roughly 50,000 images per day at its peak. ImageNet provided the critical fuel that data-hungry deep learning models needed, and its associated competition (ILSVRC) became the proving ground where breakthroughs like AlexNet demonstrated that the age of deep learning had arrived.