Deep Learning Can’t Be Trusted

Stephen Grossberg was one of my math professors when I was an undergrad at Boston University in the early 1980s. I remember it being whispered that he was using up all sorts of expensive computer time, mostly at night, doing some sort of brain simulation. This supposedly accounted for the dark circles under his eyes and his slightly disheveled appearance. As a young engineer I was instantly intrigued. I tried to read about his Adaptive Resonance Theory, but not was mostly over my head. A few years later when I took a grad level course in Artificial Intelligence I read his ART papers as part of a class. I understood the ideas, but the math, never my strong point, was over my head. A few years ago I worked with some AI folks of the new generation and mentioned ART. They weren’t familiar with it. Glad to see it’s still around. Might have to have a go at Grossbergs new book Conscious Mind, Resonant Brain: How Each Brain Makes a Mind. I remember him being an engaging lecturer. Maybe this time around I’ll understand ART a bit better. From IEEE Spectrum.

Deep Learning Can’t Be Trusted, Brain Modelling Pioneer Says Stephen Grossberg explains why his ART model is better