Artificial Intelligence Pioneer Says We Need To Start Over (axios.com)

Steve LeVine, writing for Axios: In 1986, Geoffrey Hinton co-authored a paper that, four decades later, is central to the explosion of artificial intelligence. But Hinton says his breakthrough method should be dispensed with, and a new path to AI found. Speaking with Axios on the sidelines of an AI conference in Toronto on Wednesday, Hinton, a professor emeritus at the University of Toronto and a Google researcher, said he is now "deeply suspicious" of back-propagation, the workhorse method that underlies most of the advances we are seeing in the AI field today, including the capacity to sort through photos and talk to Siri. "My view is throw it all away and start again," he said. Other scientists at the conference said back-propagation still has a core role in AI's future. But Hinton said that, to push materially ahead, entirely new methods will probably have to be invented. "Max Planck said, 'Science progresses one funeral at a time.' The future depends on some graduate student who is deeply suspicious of everything I have said."

Comments

Re:I wish they'd change terminology

[Beezlebub33]

You may be interested in OpenWorm. See: http://www.openworm.org/

They are working on simulating a worm. We can't replace individual neurons, but C. elegans is simple enough that we might be able to simulate it to the degree that we really understand it. An insect is way, way beyond what we can do now, and of course even simple vertebrates are a pipe dream. But, we're making progress. It's an open question of exactly which processes we need to simulate at what level.

As for replacing individual neurons, you'd have to know what they do in situ. Obviously, they receive signals, and they fire off other signals, but the strength of the connections change over time, the intercellular environment changes, the overall level of activity changes, they age, etc., so it's not just replacing a single neuron with a static piece of electronics; it would have to have both short and long term dynamics, and we would have to know what they are. And we don't yet.