[출처:] arXiv cs.AI

Human-Aligned Procedural Level Generation Reinforcement Learning via Text-Level-Sketch Shared Representation