🧠

Geoffrey Hinton

Original nameGeoffrey Hinton

British-Canadian Computer Scientist, "Godfather of AI," Core Founder of Deep Learning

Those Who Conceived That "a Machine Could Think" · Pioneers

The Backpropagation AlgorithmDeep LearningThe 2024 Nobel Prize in Physics

Who they are

Geoffrey Hinton (1947– ) is a British-Canadian computer scientist and cognitive psychologist who in 1986, with David Rumelhart and Ronald Williams, co-authored a paper systematically expounding the application of the "backpropagation algorithm" to the training of multi-layer neural networks, and this method of adjusting network weights by passing error gradients backward layer by layer thereafter became the core mechanism of the training of almost all deep neural networks. In the decades-long "neural network winter," when the mainstream field had largely turned to other technical paths, Hinton held always to the neural-network direction, publishing in 2006 papers on "deep belief networks" that found a more effective initialization for the multi-layer training of neural networks, widely seen as the starting point of the "deep learning revival." In 2012 the AlexNet he designed with two graduate students, Alex Krizhevsky and Ilya Sutskever, won the ImageNet image-recognition competition by an enormous margin over the runner-up, the first time deep learning proved its overwhelming advantage over the earlier mainstream methods in an authoritative public competition, thereafter triggering the whole field of AI to turn collectively to deep learning. For these foundational contributions Hinton, with Bengio and LeCun, won in 2018 the Turing Award, and in 2024 further won the Nobel Prize in Physics (shared with John Hopfield for their foundational research in machine learning with artificial neural networks). In 2023 he actively resigned his post at Google, frankly saying he resigned to be able to speak more freely about the possible long-term risks of AI, and has since continually called on all sectors of society to face these risks squarely.

Primary sourcesRumelhart, Hinton & Williams, "Learning representations by back-propagating errors" (1986)The 2024 Nobel Prize in Physics citation

Key stories

Perseverance in the Neural-Network Winter

From the late 1980s to the early 2000s, the AI field widely held the neural-network path hard to scale and limited in effect, and mainstream funding and talent turned to other statistical-learning methods such as support vector machines, but Hinton held always that the potential of neural networks was not yet fully tapped. He described those years as feeling "like defending something no one believed in," and this decades-long perseverance was at last dramatically rewarded at the 2012 ImageNet competition.

Resigning from Google Just to "Speak Freely"

When Hinton resigned from Google in 2023 he said publicly the decision was to be able to speak freely about the potential risks of AI without the constraint of an employer’s interest, and this posture of "retiring after success and turning to warning" was thereafter widely reported and discussed by the media. He has since said many times publicly that he feels both excited and uneasy about the pace of deep learning’s recent development, holding that society’s attention to the risk of loss of control such systems may bring is still far from enough.

Relationships

Mentor

🔙David Rumelhart 🖼️Fei-Fei Li　AlexNet proved the overwhelming advantage of deep learning precisely on the ImageNet dataset and competition platform Fei-Fei Li founded

Disciple

🌌Ilya Sutskever 🏆Demis Hassabis

Ally

👁️Yann LeCun　Named with LeCun and Bengio the "three giants of deep learning," together winning the Turing Award in 2018 🧭Yoshua Bengio　Named with LeCun and Hinton the "three giants of deep learning," together winning the Turing Award in 2018 ⚡Terrence Sejnowski　The two co-proposed the Boltzmann machine, an early neural-network model

Nemesis

⏳Jürgen Schmidhuber　Schmidhuber has publicly accused the mainstream narrative of deep learning many times of underrating the contribution of his own and his team’s earlier research, the two having long public disputes over several priority questions

Echoes today

Below are how modern works borrow or reinterpret this name or story — not the original material. The two differ, so keep them apart.

The "godfather of AI" resigns to warn of riskHinton’s move of resigning from Google in 2023 and thereafter continuing to warn publicly of AI’s potential risk was widely reported by the media as a landmark event of "retiring at the peak, turning to warn."

Appears in

The Perceptron, the Winter, and the Revival of ConnectionismThe Rediscovery of the Backpropagation Algorithm The Deep-Learning Revival and the Shock of ImageNetPersistence Through the "Neural-Network Winter" and the Shock of AlexNet

← Back to Figures