3 ms·DRL is too fragile to be realistically useful while supervised learning is not.by tanilama 9y agoDRL is too fragile to be realistically useful while supervised learning is not.