Summary of Mastering Chinese Chess Ai (xiangqi) Without Search, by Yu Chen et al.
Mastering Chinese Chess AI (Xiangqi) Without Search
by Yu Chen, Juntong Lin, Zhichao Shu
First submitted to arxiv on: 7 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research paper presents a high-performance Chinese Chess AI that can compete at an elite level without relying on search algorithms. The AI achieves exceptional performance, exceeding systems based on Monte Carlo Tree Search (MCTS) by over 1,000 times and those using AlphaBeta pruning by more than 100 times. The training system combines supervised learning to create a human-like AI with reinforcement learning to further enhance its strength. The study also explores various techniques, including the use of Transformer architecture, possible moves as features, selective opponent pools, and value estimation with cutoff (VECT), which improve the training process. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper makes a Chinese Chess AI that can play really well without using special searching methods. This AI is very strong, beating most other systems by a lot! To make this AI work, researchers combined two types of learning: one to create a basic AI and another to make it even stronger. They also tried different approaches to see what works best, such as using new computer architectures or selecting the right opponents. |
Keywords
» Artificial intelligence » Pruning » Reinforcement learning » Supervised » Transformer
