Melanie Mitchell, a professor of complexity at the Santa Fe Institute and a professor of computer science at Portland State University, acknowledges the power of black-box deep learning networks. She also ponders and wonders that artificial intelligence research would benefit most from getting back to its roots and exchanging more ideas with researchers into cognition in living brains.
This week, she spoke with host Steven Strogatz about the challenges and the difficulties in building general intelligence, why we should think about the road rage of self-driving cars, and why AIs might need good parents.
Doug Hofstadter, whom she had met with years ago had developed the program called Copycat. Melanie asserted that the idea behind the program was to do the same thing but in a different way and a different situation. The computer program cannot be advised to perform complex analogies and the aim was to make the general intelligence simpler which can be easily instilled into the systems of gigantic enterprises. Melanie’s point is that we’re interested in this broad problem of how you can probe analogies inside an artificial intelligence and thereby get the essence of what meaning it has and will have in the real-life scenario.