AlexNet wins ImageNet

AlexNet, a convolutional neural network developed by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton at the University of Toronto, achieved a groundbreaking 15.3% error rate in the ImageNet competition, significantly outperforming the second-place entry (26.2%). It utilized NVIDIA GPUs and custom CUDA code for its training.

Significance

This victory validated the power of deep convolutional neural networks and GPU acceleration for computer vision tasks, sparking a massive surge of interest and investment in deep learning and demonstrating the critical need for efficient GPU programming tools like cuDNN.