E. Korkmaz, “Deep Reinforcement Learning Policies Learn Shared Adversarial Features across MDPs”, AAAI, vol. 36, no. 7, pp. 7229-7238, Jun. 2022.