Reinforcement learning!

[Bartleby2718]

• Useful Links

  • QMaze (Very useful)
  • TensorFlow docs
  • Keras docs
  • Others: https://medium.com/@dhruvp/how-to-write-a-neural-network-to-play-pong-from-scratch-956b57d4f6e0 https://www.analyticsvidhya.com/blog/2017/01/introduction-to-reinforcement-learning-implementation/

• Memo

  • AI로 할 것 : joker value, joker position, deck choice, action
  • It would definitely be advantageous to know the opponent's strategy, but you can only make an educated guess via induction, but induction might not always be correct. But then there are a lot of papers written about "opponent modeling in multi-agent reinforcement learning." Hmm...