5

I read through the publication Mastering the game of Go without Human Knowledge. It doesn't seem to use GANs, just a new form of search and reinforcement learning.

nbro
  • 39,006
  • 12
  • 98
  • 176
dougvk
  • 163
  • 6

1 Answers1

10

No, GANs are not used. It's reinforcement learning at what it does best. The tree search is an interesting addition and assists with navigating the sheer scale of the game.

Although the agent was playing itself to become better, there wasn't 2 separate networks (generator and discriminator). The agent learned through RL and didn't have the error calculations that generators and discriminatory have.

dougvk
  • 163
  • 6
Jaden Travnik
  • 3,767
  • 1
  • 16
  • 35
  • I found this article helpful re: what the search algorithm is: https://medium.com/@bramcohen/alphazeros-real-secret-sauce-is-mcts-b6f238d95374 – dougvk Dec 24 '17 at 13:27