Contrary to claims that I somehow "dismiss" the idea reasoning in DL systems, I've long listed 3 main challenges to AI in my talks of the last several years, one of which is "learning to reason, in ways that are compatible with gradient-based learning" https://t.co/Oxc7dR8jF4 pic.twitter.com/VoCdb8h3mP
— Yann LeCun (@ylecun) May 25, 2022
Perhaps there's an idea or two in the paper I'm now working on, Relational Nets Over Attractors, A Primer: Part 1, Basics.
