让游戏关卡生成更听指令，支持多目标复杂命令。
Multi-Objective Instruction-Aware Representation Learning in Procedural Content Generation RL
- 用句子嵌入构建多目标条件空间，提升文本理解力
- 多标签分类与多头回归联合训练，控属性提升13.8%
- 适合需要精细语义控制的生成任务开发者
生成模型越来越依赖自然语言作为内容生成的控制方式。然而，现有指导式强化学习在程序化内容生成（IPCGRL）中难以充分挖掘文本输入的表达潜力，尤其在复杂多目标指令下，导致可控性受限。为此，我们提出MIPCGRL，一种面向指令内容生成的多目标表征学习方法，将句子嵌入作为条件输入。该方法通过多标签分类与多头回归网络联合训练，构建有效的多目标嵌入空间。实验表明，在多目标指令下，该方法可实现最高13.8%的可控性提升，显著增强内容生成的表达力与灵活性。
原文摘要 · Abstract (English)
Recent advancements in generative modeling emphasize the importance of natural language as a highly expressive and accessible modality for controlling content generation. However, existing instructed reinforcement learning for procedural content generation (IPCGRL) method often struggle to leverage the expressive richness of textual input, especially under complex, multi-objective instructions, leading to limited controllability. To address this problem, we propose \textit{MIPCGRL}, a multi-objective representation learning method for instructed content generators, which incorporates sentence embeddings as conditions. MIPCGRL effectively trains a multi-objective embedding space by incorporating multi-label classification and multi-head regression networks. Experimental results show that the proposed method achieves up to a 13.8\% improvement in controllability with multi-objective instructions. The ability to process complex instructions enables more expressive and flexible content generation.
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。