Hacker News Reader

I used to believe as you do until AlphaZero learn to play Go and Chess by itself and discovering new strategies in the process.

By using tensor calculus.

Can you provide link to paper explaining the use of tensor calculus in AlphaZero?