1 paper across 1 session
We achieve sample-efficient visual policy learning under the supervisions of only expert videos, without accessing rewards or expert actions.
