Alex Krizhevsky is a Canadiancomputer scientist most noted for his work onartificial neural networks anddeep learning. In 2012, Krizhevsky,Ilya Sutskever and their PhD advisorGeoffrey Hinton, at theUniversity of Toronto,[1] developed a powerful visual-recognition networkAlexNet using only twoGeForce-brandedGPU cards.[2] This revolutionized research in neural networks. Previously neural networks were trained onCPUs. The transition to GPUs opened the way to the development of advancedAI models.[2]
Motivated by Sutskever and inspired by Hinton, Krizhevsky developed AlexNet to expand the limits in image recognition and classification. Building onConvolutional Neural Networks and Sutskever’sDeep Neural Network approach of deepening the neural layers far beyond the convention of the time - as well as addingDropout for training resilience -AlexNet won theImageNet challenge in 2012. The team presented their paper for AlexNet[3] atNeurIPS (NIPS) 2012.
Shortly after AlexNet’s debut, Krizhevsky and Sutskever sold their startup, DNN Research Inc., toGoogle. Krizhevsky left Google in September 2017 after losing interest in the work, to work at the company Dessa in support of new deep-learning techniques.[1] Many of his numerous papers onmachine learning andcomputer vision are frequently cited by other researchers.[4] He is also the main author of theCIFAR-10 and CIFAR-100 datasets.[5][6]
AlexNet is widely credited with igniting thedeep learning revolution. Its success demonstrated the effectiveness of deep neural networks trained on GPUs, leading to rapid progress across multiple domains of artificial intelligence beyond computer vision. The techniques and momentum generated by AlexNet helped shape the development of modern natural language processing models, including large-scaletransformer-based models such asBERT andGPT, which power tools likeChatGPT.[7]
