Reinforcement learning in linear embedding space unlocks generalizable control across soft robot configurations figure
AlphaXiv 中文概览(可滚动查看)