AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Artificial Intelligence Deep Learning
Deep learning is solving the same kind of problems that humans do, but doing it apparently radically differently. And then there is the fact that we need to train the network on 100,000 or a million images where the human brain needs nothing of the sort. You can game the system by altering pixels invisibly to make it think that a horse is an ostrich. That's fundamentally at odds with what we think about how humans do this. So, what are your thoughts around this kind of dichotomy?