Does generator + Goldilocks critic produce measurably more learning progress than uniform random goal sampling at matched interaction budget?
reinforcement-learning intrinsic-motivation curriculum-learning learning-progress goal-conditioned-rl negative-result
-
Updated
Sep 17, 2026 - Python