1 paper
Yunhao Zhao, Zhenyang Ni, Haoyang Chen +2
Modern vision-language-action (VLA) policies have acquired broad manipulation skills, but typically generate each action chunk from the current observation or a short fixed-length…