AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Biggest Challenge for Neural Networks
Markis: I don't see any reason to suppose thatour systems are still going to get things wrong, and badly wrong intensely. The big question is, would it get things, err badly wrong less often a human, if it does, do better than a human in most of the time? And perhaps that's good enough. What you were saying about the monte carlo tree search is fascinating, markis. I guess we're talking about what the future is and neuro symbolic methods. There just ways of incoding discreet information into a neural network.