1 paper
Zetao Cai, Yaping Li, Yiqun Wang +8
Robot demonstrations are expensive to collect and often provide limited distributional coverage of task variations. Human videos offer a low-cost source of complementary manipulati…