A neural network based Legend of the 5 Rings LCG AI written in C++
This AI is inspired by AlphaGo Zero. It combines Monte Carlo tree search with deep neural network reinforcement learning. It uses only self play to improve itself.
The state/action space for L5R is very large especially considering the entire card pool. Also hidden information/randomness is still a challenge to take into consideration.
As a simple initial proof of concept the initial version will play with the following limitations.
- no hidden information (hand exposed, and no deck shuffling)
- turn based bidding starting with first player
- lion and crane starter decks only
- I am an L5R player myself looking to better my own gameplay and deckbuilding
- I have an interest in programming and machine learning
Once the initial engine is complete and can successfully learn to play with these limitations I would like to start to experiment and evolve it with more features.
- Hidden information
- Supporting different clan matchups
- Deckbuilding and deck strength evalutation
- Increase to support card pool and evolve with the meta
- Add a GUI to the engine for easy human play