sep 30, 2012 - ImageNet Classification with Deep Convolutional Neural Networks

Description:

Krizhevsky, Sutskever, and Hinton introduce AlexNet, a deep CNN (convolutional neural network) that dramatically advances image classification performance. Trained on GPUs - a novel choice at this time - AlexNet won the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) by a wide margin, reducing top-5 error rates from ~26% to 15%. Its architecture includes multiple convolutional and fully connected layers, ReLU activations, dropout for regularization, and data augmentation techniques. This success shows that deep networks, when paired with large datasets and GPU acceleration, can far outperform traditional vision methods. The result triggers a surge of interest in deep learning for computer vision and is widely credited with launching the modern AI boom.

Added to timeline:

Date:

Images: