2000 character limit reached
Fast and Knowledge-Free Deep Learning for General Game Playing (Student Abstract) (2312.14121v1)
Published 21 Dec 2023 in cs.AI
Abstract: We develop a method of adapting the AlphaZero model to General Game Playing (GGP) that focuses on faster model generation and requires less knowledge to be extracted from the game rules. The dataset generation uses MCTS playing instead of self-play; only the value network is used, and attention layers replace the convolutional ones. This allows us to abandon any assumptions about the action space and board topology. We implement the method within the Regular Boardgames GGP system and show that we can build models outperforming the UCT baseline for most games efficiently.
- A Survey of Monte Carlo Tree Search Methods. IEEE Transactions on Computational Intelligence and AI in Games, 4(1): 1–43.
- Minimax Strikes Back. In AAMAS, 1923–1931.
- General Game Playing: Overview of the AAAI Competition. AI Magazine, 26: 62–72.
- Deep Reinforcement Learning for General Game Playing. AAAI, 34(02): 1701–1708.
- Efficient Reasoning in Regular Boardgames. In IEEE Conference on Games, 455–462.
- Split Moves for Monte-Carlo Tree Search. AAAI, 36(9): 10247–10255.
- Regular Boardgames. AAAI, 33(1): 1699–1706.
- Rosin, C. D. 2011. Multi-armed bandits with episode context. Annals of Mathematics and Artificial Intelligence, 61(3): 203–230.
- A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play. Science, 362(6419): 1140–1144.
- Deep learning for general game playing with ludii and polygames. ICGA Journal, 43(3): 146–161.
- Attention is all you need. NeurIPS, 30.