Geoffrey Hinton talking with his hands

Geoffrey Hinton. (Daniel Ehrenworth photos)

Innovation | From Pivot Magazine

Machine Man

How Geoffrey Hinton, the godfather of artificial intelligence, changed the way we do business

A Facebook IconFacebook A Twitter IconTwitter A Linkedin IconLinkedin An Email IconEmail

Geoffrey Hinton, a lanky, bemused 70-year-old deserves a good gloat right about now. For three decades, he toiled on the fringes of artificial intelligence in an unadorned office at the University of Toronto, garnering scoffs from his peers. Descended from a long line of British scientists—his great great-grandfather was the computer pioneer George Boole, of the “Boolean search”—Hinton never lost faith in “machine learning,” the idea that computers could learn like humans, using intuition rather than logic. He experimented with neural networks modelled on the human brain, where information pings between synapses, building layers of intelligence. Meanwhile, most everyone else in AI threw their weight behind logic-based programming.

Then around 2012, it turned out that he was right all along. Computers became powerful enough to execute Hinton’s theories. Tech giants like Facebook, Apple, OpenText and Uber adopted Hinton’s pioneering version of AI. His former students from U of T became the hottest recruits in Silicon Valley. And machine learning became the path to change our lives: think of companies like Sage and Xero and their AI-based accounting software. Think selfdriving cars, automated transcription and an app that will diagnose cancerous lesions. Every time your phone nudges you to hit the quick response to an email (“See you soon!”), that’s neural networks. Every time an accounting app suggests where to allocate a particular transaction, that’s Hinton’s work.

Analysts say that by 2020, machine learning will take on the most tedious tasks of bookkeeping, which means auditors will have access to data that’s more complete, timely and accurate than ever before. So this cutting-edge technology might deliver what humans have always coveted: more free time.

Meanwhile, Hinton continues his dogged research as an engineering fellow at Google and the chief scientist at the just-launched Vector Institute, a $180-million AI research centre in downtown Toronto. The institute is a public-private hybrid effort to transform Toronto into the AI capital of the world. As the doors opened this winter, Hinton announced a new game-changing breakthrough: “capsule networks,” which will revolutionize a machine’s ability to recognize images. Vindication never sleeps.