World Models as Adversaries: Multi-Agent Self-Play Fine-Tuning for Robust Motion Planning figure
AlphaXiv 中文概览(可滚动查看)