Geoffrey E. Hinton

Professor of Computer Science and Distinguished Researcher at Google. Co-inventor of the Boltzmann Machine learning algorithm. Inventor of variational Bayesian learning. Pioneered the use of back-propagation for learning distributed representations of semantic knowledge.  Showed that unsupervised Boltzmann Machine learning could be used to initialize the weights of deep neural networks. Showed that back-propagation in deep neural networks outperformed existing methods for speech recognition and image classification.