Geoffrey Hinton is usually credited: his 2006 deep belief net paper with Osindero and Teh made the term popular. Strictly, he didn't coin it. Rina Dechter used "deep learning" in 1986, and Igor Aizenberg and colleagues applied it to neural networks in 2000.
Card 1 of 259. Answer: Geoffrey Hinton is usually credited: his 2006 deep belief net paper with Osindero and Teh made the term popular. Strictly, he didn't coin it. Rina Dechter used "deep learning" in 1986, and Igor Aizenberg and colleagues applied it to neural networks in 2000.
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
Geoffrey Hinton is the usual answer, but "popularized" is more accurate than "coined". In 2006 Hinton, Simon Osindero and Yee-Whye Teh published "A fast learning algorithm for deep belief nets" in Neural Computation, the paper usually credited with making "deep learning" the name for training neural networks with many layers.
The phrase itself is older. Rina Dechter used "deep learning" in a 1986 AAAI paper on constraint satisfaction, which was not about neural networks. Igor Aizenberg and colleagues applied it to neural networks in a 2000 book on multi-valued and universal binary neurons. Hinton himself has said that the term's current popularity started with the 2006 paper.
What was new in 2006 was a method: pretrain a deep network one layer at a time as a stack of restricted Boltzmann machines, then fine-tune it. That made many-layer networks trainable when many researchers had given up on them. Hinton later shared the 2018 Turing Award with Yoshua Bengio and Yann LeCun.
