Geoffrey Hinton

Co-authored the paper that made backpropagation widely known, and led the deep learning revival.

Biography

A British-Canadian cognitive psychologist and computer scientist who continued working on neural networks through the period when the field had largely moved on. He co-authored the 1986 work that brought backpropagation to broad attention, and his group’s 2012 entry to the ImageNet competition is widely treated as the moment deep learning became dominant in computer vision.

Key contributions

• Co-authored Learning Internal Representations by Error Propagation (1986) with Rumelhart and Williams.
• Work on Boltzmann machines and unsupervised pretraining of deep networks.
• Led the group whose 2012 ImageNet result established deep convolutional networks in vision.

Impact

The attribution for backpropagation is genuinely layered: equivalent methods had been derived independently by earlier authors, and the 1986 paper is best understood as the point at which the technique became widely known rather than first invented. Chollet observes that convolutional networks and backpropagation were both well understood by 1990, and that what changed after 2012 was the availability of data and compute.