#Neural-Networks

News Feature: What are the limits of deep learning?

PNAS profile of 2019 surveys where deep learning breaks: adversarial spoofing, data inefficiency, opacity, and the missing common sense that symbolic AI once tried to hand-code.