arXiv：cs.LG（机器学习，全量分类）· Debosmita Bhaumik, Julian Togelius, Georgios N. Yannakakis, Ahmed Khalifa·· 18 小时前AI 评分29
用 WFC 学习局部约束来约束强化学习内容生成器 PCGRL
Learning Local Constraints for Reinforcement-Learned Content Generators
AI 导读
研究者将 Wave Function Collapse（WFC）学到的局部约束用于限制 PCGRL 生成器的动作空间，使强化学习生成器在满足局部约束的同时优化可玩性等全局属性。该方法对超参数调优较敏感，但最佳生成器能产出视觉满意且可玩的 Lode Runner 类解谜平台关卡。
正文
Abstract:Constraint-based game content generators that learn local constraints from existing content, such as Wave Function Collapse (WFC), can generate visually satisfying game levels but face challenges in optimizing global properties, such as playability. On the other hand, reinforcement-learning-trained generators can optimize global properties---because such properties can easily be included in reward functions---but the results can be visually dissatisfying. In this paper, we explore ways to combine these methods. Specifically, we constrain the action space of a PCGRL generator with constraints learned by WFC, effectively allowing the PCGRL generator to achieve global properties while being forced to adhere to local constraints. To better analyze how this hybrid content generation method operates, we vary the number and type of inputs, and we test whether to randomly collapse the starting state and exclude rare patterns. While the method is sensitive to hyperparameter tuning, the best of our trained generators produce visually satisfying and playable puzzle-platform game levels---such as Lode Runner levels---with desired global properties.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.13570 [cs.AI] |
| (or arXiv:2605.13570v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.13570 arXiv-issued DOI via DataCite |
Submission history
From: Debosmita Bhaumik [view email]
[v1]
Wed, 13 May 2026 14:07:10 UTC (2,191 KB)
[v2]
Thu, 1 Oct 2026 03:08:31 UTC (2,242 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org