Machine Learning Blog ML@CMU Carnegie Mellon University

The figure above illustrates the method:(a) Goal-conditioned RL often fails to reach distant goals, but can successfully reach the goal if starting nearby (inside the green region).