5 ms·

I believe AlphaGo Zero used reinforcement learning. I would say that's quite an impressive application of it.