用JSON特征训练通用游戏价值函数，提升智能体决策能力
Game-Agnostic Value Functions through Automatic JSON Feature Extraction
- 基于JSON特征提取构建通用游戏价值函数框架
- 随机森林特征选择使性能显著提升，优于传统方法
- 适合跨游戏智能体训练与强化学习研究者使用
JSON Bag-of-Tokens（JSON-Bag）是一种近期提出的通用方法，通过将游戏轨迹的JSON描述分词，实现对游戏状态的统一表示。本文提出JSON-Bag VF，一种基于JSON-Bag原型的、适用于各类桌游的价值函数训练方法。该方法可结合基于随机森林的特征选择及针对游戏阶段特异性特征的选择策略进行优化。我们在六种桌游上，针对不同原型分词方式与特征选择组合，评估了基于一步前瞻（One-step-look-ahead）的JSON-Bag OSLA模型表现。实验结果表明，多数游戏中JSON-Bag OSLA均优于基线方法。分析显示，特征选择对模型性能影响最大，甚至超过原型分词方式本身的影响。
原文摘要 · Abstract (English)
JSON Bag-of-Tokens (JSON-Bag) is a recently proposed method to generically represent game trajectories by tokenizing their JSON descriptions. We introduce JSON-Bag VF, a game-agnostic approach to training value functions for game-playing agents using JSON-Bag prototypes. We show that this approach can be enhanced with Random Forest-based feature selection and a method to select game-stage-specific features. We evaluate JSON-Bag VF with One-step-look-ahead (JSON-Bag OSLA) on six tabletop games over different combinations of prototype-tokenization and feature selections. JSON-Bag OSLA outperforms baseline OSLA agents in most games. Our analysis also shows that feature selection significantly improves JSON-Bag VF and that feature selection is the most important factor in JSON-Bag VF performance, over prototype-tokenization.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。