Implement reinforcement learning roadmap

[0xsuu]

Implement Route:

• Traditional Tabular Method:
  • Monte Carlo Value Iteration on Tiny Mahjong ->
  • Q-learning on Tiny Mahjong
• Learn the rules, learn the strategy:
  • Tiny Mahjong ->
  • Simple Mahjong Random ->
  • Gomoku Random ->
  • Simple Mahjong Greedy ->
  • Gomoku Beginner ->
  • Japanese Mahjong
• Learn to react:
  • Breakout ->
  • Snake
• Self-playing:
  • Simple Mahjong Duel/Full ->
  • Japanese Mahjong
• Searching:
  • Tiny Mahjong (break the optimal) ->
  • Simple Mahjong ->
  • Japanese Mahjong
• Value prediction:
  • Tiny Mahjong ->
  • Simple Mahjong ->
  • Gomoku ->
  • Japanese Mahjong