ImageNet and the Deep Learning Revolution

In October 2012, a deep convolutional neural network called AlexNet, built by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton at the University of Toronto, won the ImageNet Large Scale Visual Recognition Challenge with a top-5 error rate of 15.3% — nearly half the 26.2% of the runner-up. ImageNet, a dataset of over 14 million labeled images created by Fei-Fei Li's team at Stanford, had provided the fuel; NVIDIA GPUs provided the computational power; and Hinton's decades of work on backpropagation provided the architecture. The result electrified the field. Within months, major tech companies were hiring deep learning researchers, and the modern AI boom began. The "deep learning revolution" would soon spread from vision to speech recognition, translation, and eventually language generation.