Hands-On Machine Learning for Algorithmic Trading

AlexNet – putting CNN on the map (2012)

AlexNet, developed by Alex Krizhevsky, Ilya Sutskever, and Geoff Hinton at the University of Toronto, significantly outperformed the runner-up at the 2012 ILSVRC (top 5 error of 16% versus 26%). The performance breakthrough achieved by AlexNet triggered a renaissance of ML research, and put deep learning for computer vision firmly on the global technology map.

The AlexNet architecture is similar to LeNet, but much deeper and wider, and included convolutions stacked on top of each other rather than combining each convolution with a pooling stage.

It discovered the importance of depth, and successfully used dropout for regularization and ReLU as efficient non-linear transformations. It also employed data ...