Mind Matters News and Analysis on Natural and Artificial Intelligence
George Montañez

What is learning anyway?

Machine learning specialist George Montañez reflects on the question in a video excerpt from the CNAI gala
George Montañez chatting with Michael Behe at the Walter Bradley Center launch in July

The official launch of the Walter Bradley Center for Natural and Artificial Intelligence (July 11, 2018) featured an evening of livestreamed discussion around the issue, “Will the Machines Take Over?: Human Uniqueness in the Age of Smart Machines.”

Microsoft data scientist and machine learning specialist George Montañez addressed the question of “What is learning anyway?”

Alan Turing, who was the luminary father of computer science in many ways, presented the question, “Can machines think?” And he felt that that question was very vague and very hard to answer and so he suggested let’s try a a simpler question: Can we make machines that can reliably fool human judges into thinking that the machine a human? So, in other words, I can paraphrase this: Can we make approximations that are so close to ourselves that the fact that they are approximations no longer matters? And I think this is an important more concrete question. … I feel that this is an open question. … Even if machines can never fully replicate human beings, for a lot of things, we do not need all of human ability to do them. For the Turing Machine you do not need the full panoply of human ability to have this machine work; you just need a finite state machine machine. 6:12ff min

See also: How Humans Can Thrive in a World of Increasing Automation (William Dembski)


What is learning anyway?