3 ms·

I wonder if it is possible to achieve super-human play purely using reinforcement learning to train deep neural networks.