Mind Matters News and Analysis on Natural and Artificial Intelligence

Deep Learning won’t solve AI

AlphaGo pioneer: We need “another dozen or half-a-dozen breakthroughs”
Share
Facebook
Twitter
googleplus Google+
arroba Email
Demis Hassabis
Demis Hassabis

There are doubtless many truths about AI but artificial intelligence pioneer Demis Hassabis, part of the AlphaGo team, offers three, one of which is that Deep Learning “is not enough to crack” general AI:

Creating a machine with a general intelligence similar to our own will require a wider range of technologies than the deep-learning systems that have powered many recent breakthroughs.

“Deep learning is an amazing technology and hugely useful in itself, but in my opinion it’s definitely not enough to solve AI, [not] by a long shot,” he said.

“I would regard it as one component, maybe with another dozen or half-a-dozen breakthroughs we’re going to need like that. There’s a lot more innovations that’s required…

AlphaGo logo

“AlphaGo doesn’t understand language but we would like them to build up to this symbolic level of reasoning — maths, language, and logic. So that’s a big part of our work,” he said, adding DeepMind is also working on how to make learning more efficient, in order to reduce the huge volume of data needed to train deep learning systems today. Nick Heath, “Google DeepMind founder Demis Hassabis: Three truths about AI” at TechRepublic

One difficulty Hassabis’s team may face is the general imprecision of language, whose intentionally fuzzy boundaries are constantly under negotiation and revision. In many cases, correct interpretation requires emotional intelligence.

Hat tip: Eric Holloway

See also: Artificial intelligence is impossible Eric Holloway

and

See also: Could one single machine invent everything (Eric Holloway)

Demis Hassabis:

AlphaGo: