Abstract: Despite substantial progress in deep reinforcement learning (DRL), a systematic characterization of DRL agents’ robustness to unexpected events in the environment is relatively understudied.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback