For the last 4 decades, Geoff Hinton has been working on AI. His quest to build complex artificial intelligence systems began in high school, when one of his friends convinced him that the brain worked like a hologram. Even back in 1960s Britain, Hinton was intrigued by the idea that the brain could store images in much the same way as those 3D holographic images. Except, rather than storing them in one location, it spreads them across a neural network. Since then, Hinton has been working to try and replicate the way that the brain works, in machines. Let’s meet the godfather of modern AI.
Early work: Boy wonders from the very beginning
Hinton was deeply fascinated by his friend’s revelation. He says, “I got very excited about that idea.That was the first time I got really into how the brain might work.” Hinton finished high school, and went on to explore neural networks at the University of Edinburgh and Cambridge University. By the early 80s, he was helping to mimic the brain using computers, to create a pure form of AI that we call “deep learning.”
For three decades, deep learning was somewhat of an underground movement in academia. Now, Hinton and his group of colleagues, have the attention of some of the biggest names in technology.
After spending some time honing his ideas as a professor and researcher at the University of Toronto, Canada, he now works part-time for Google. There, he’s using deep learning to improve various online tools like image tagging, and voice recognition.
Hinton gathers data and reacts to it
Want to meet the godfather of modern AI? You need an understanding of his work. Hinton is fascinated by brains, and part of his work revolves around understanding them. He doesn’t have all of the answers, but he’s a lot closer to them thanks to his work with building artificial neural networks. Hinton remarks, “I get very excited when we discover a way of making neural networks better – and when that’s closely related to how the brain works.”
In his world, a neural network is software that operates on multiple different levels. Alongside his colleagues, he builds artificial neurons that are modeled on the columns of neurons you’d find in the brain’s cortex – which deals with vision and language.
These AI neural networks can gather data, and react to it – building an understanding of what something looks or sounds like. They’re getting even better at figuring out what words mean when you string them together, and they can do all of this without the need for a human to step in and provide labels for everything.
These neural networks are extremely fast, efficient, and adept at many complex tasks. They’re also about 30 years in the making.
Hinton, a revolutionary in the field
Hinton hasn’t sat down in 12 years thanks to a back condition, so he walks everywhere. In some ways, that feels relevant to the level of persistence that he’s shown in the field of AI. Hinton was building and researching neural networks long before the rest of the world caught on, and that’s why you can now meet the godfather of modern AI. Without his work, which spans 4 decades, AI research would likely be in a very different place – that’s quite remarkable.