Communications of the ACM
Polanyi's Revenge and AI's New Romance with Tacit Knowledge
Credit: Getty Images
In his 2019 Turing Award Lecture, Geoff Hinton talks about two approaches to make computers intelligent. One he dubs—tongue firmly in cheek—"Intelligent Design" (or giving task-specific knowledge to the computers) and the other, his favored one, "Learning" where we only provide examples to the computers and let them learn. Hinton's not-so-subtle message is that the "deep learning revolution" shows the only true way is the second.
Hinton is of course reinforcing the AI Zeitgeist, if only in a doctrinal form. Artificial intelligence technology has captured popular imagination of late, thanks in large part to the impressive feats in perceptual intelligence—including learning to recognize images, voice, and rudimentary language—and bringing fruits of those advances to everyone via their smartphones and personal digital accessories. Most of these advances did indeed come from "learning" approaches, but it is important to understand the advances have come in spheres of knowledge that are "tacit"—although we can recognize faces and objects, we have no way of articulating this knowledge explicitly. The "intelligent design" approach fails for these tasks because we really do not have conscious theories for such tacit knowledge tasks. But, what of tasks and domains—especially those we designed—for which we do have explicit knowledge? Is it forbidden to give that knowledge to AI systems?
Comments
Joseph Bedard
Thank you for this article. You have elegantly, concretely and thoroughly described a problem in deep learning systems that has been subconsciously nagging me for the past 4 years.
Huan Liu
Bravo!
Shai Ben-David
This is an important, timely and clearly written "back to common sense" article. There are no miracles in AI and machine learning, and the No Free Lunch *theorem*, stating that learning *requires* prior knowledge, has not been, and cannot be, refuted by either `proofs by examples' (deep learning successes) or by `proof by authority' ("Hinton said").
Amit Sheth
Truly enjoyed reading this perspective. It is a #MustRead for many in AI as knowledge is increasingly playing a larger role to address the deficiencies of pure data-driven and statistical AI approaches. This perspective inspired me to write down a few complementary views: https://www.linkedin.com/pulse/mini-community-reflection-raos-article-romance-tacit-knowledge-sheth/
Displaying all 4 comments