Skip to content

amndazhang/BetaBot

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 

Repository files navigation

BetaBot

AlphaGo's beta counterpart.

game board state (encoded for neural networks): tracked move-by-move with game state tree

tree search: makes predictions on sequence of decisions

neural network: extract game-relevant features w/ CNNs, pool layers, softmax activations on last layer, categorical cross-entropy as loss func

deep learning: build data with Smart Game Format, encode important properties (like num of liberties) in feature planes

reinforcement learning: training agent to learn from experience w/ states, actions, rewards; further collect data from self-play

policy gradients:

value methods:

actor-critic methods:

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages