Cumulative Reward on Qbert
22,840.56Cumulative RewardNo Attack
Evaluation Results
| Method | Links | |
|---|---|---|
| No AttackAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 22,840.56 | |
| No AttackAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 22,840.56 | |
| No AttackAttack Setting=White-box, Target RL Policy=A3C2026.02 | 21,840.35 | |
| No AttackAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 21,840.35 | |
| No AttackAttack Setting=White-box, Target RL Policy=DQN2026.02 | 17,890.93 | |
| No AttackAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 17,890.93 | |
| No AttackAttack Setting=White-box, Target RL Policy=PPO2026.02 | 2,458.68 | |
| No AttackAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 2,458.68 | |
| EDGEAttack Setting=White-box, Target RL Policy=PPO2026.02 | 1,246.21 | |
| EDGEAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 1,246.21 | |
| SkipAttack Setting=White-box, Target RL Policy=PPO2026.02 | 680.34 | |
| SkipAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 680.34 | |
| SkipAttack Setting=White-box, Target RL Policy=DQN2026.02 | 678.35 | |
| SkipAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 678.35 | |
| SkipAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 667.32 | |
| SkipAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 667.32 | |
| S-TAttack Setting=White-box, Target RL Policy=A3C2026.02 | 569.38 | |
| S-TAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 569.38 | |
| S-TAttack Setting=White-box, Target RL Policy=PPO2026.02 | 568.23 | |
| S-TAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 568.23 | |
| S-TAttack Setting=White-box, Target RL Policy=DQN2026.02 | 543.44 | |
| S-TAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 543.44 | |
| SkipAttack Setting=White-box, Target RL Policy=A3C2026.02 | 543.27 | |
| SkipAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 543.27 | |
| S-TAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 510.35 | |
| S-TAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 510.35 | |
| EDGEAttack Setting=White-box, Target RL Policy=A3C2026.02 | 253.96 | |
| EDGEAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 253.96 | |
| EDGEAttack Setting=White-box, Target RL Policy=DQN2026.02 | 235.68 | |
| EDGEAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 235.68 | |
| FGSMAttack Setting=White-box, Target RL Policy=PPO2026.02 | 210.32 | |
| FGSMAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 210.32 | |
| FGSMAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 205.33 | |
| FGSMAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 205.33 | |
| EDGEAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 199.84 | |
| EDGEAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 199.84 | |
| FGSMAttack Setting=White-box, Target RL Policy=A3C2026.02 | 189.34 | |
| FGSMAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 189.34 | |
| PA-ADAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 175.62 | |
| PA-ADAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 175.62 | |
| FGSMAttack Setting=White-box, Target RL Policy=DQN2026.02 | 165.43 | |
| FGSMAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 165.43 | |
| PA-ADAttack Setting=White-box, Target RL Policy=DQN2026.02 | 152.34 | |
| PA-ADAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 152.34 | |
| PA-ADAttack Setting=White-box, Target RL Policy=A3C2026.02 | 148.92 | |
| PA-ADAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 148.92 | |
| AdvRL-GANAttack Setting=White-box, Target RL Policy=A3C2026.02 | 139.84 | |
| AdvRL-GANAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 139.84 | |
| AdvRL-GANAttack Setting=White-box, Target RL Policy=PPO2026.02 | 130.58 | |
| AdvRL-GANAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 130.58 | |
| AdvRL-GANAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 129.83 | |
| AdvRL-GANAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 129.83 | |
| PIAAttack Setting=White-box, Target RL Policy=A3C2026.02 | 129.46 | |
| PIAAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 129.46 | |
| AdvRL-GANAttack Setting=White-box, Target RL Policy=DQN2026.02 | 124.82 | |
| AdvRL-GANAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 124.82 | |
| PIAAttack Setting=White-box, Target RL Policy=PPO2026.02 | 117.43 | |
| PIAAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 117.43 | |
| PIAAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 109.49 | |
| PIAAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 109.49 | |
| PA-ADAttack Setting=White-box, Target RL Policy=PPO2026.02 | 101.23 | |
| PA-ADAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 101.23 | |
| PIAAttack Setting=White-box, Target RL Policy=DQN2026.02 | 98.93 | |
| PIAAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 98.93 | |
| TSGEAttack Setting=White-box, Target RL Policy=PPO2026.02 | 94.35 | |
| TSGEAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 94.35 | |
| TSGEAttack Setting=White-box, Target RL Policy=A3C2026.02 | 83.98 | |
| TSGEAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 83.98 | |
| AATAttack Setting=White-box, Target RL Policy=A3C2026.02 | 60.22 | |
| AATAttack Setting=Black-box, Target RL Policy=A3C2026.02 | 60.22 | |
| AATAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 58.42 | |
| AATAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 58.42 | |
| TSGEAttack Setting=White-box, Target RL Policy=TRPO2026.02 | 58.01 | |
| TSGEAttack Setting=Black-box, Target RL Policy=TRPO2026.02 | 58.01 | |
| AATAttack Setting=White-box, Target RL Policy=PPO2026.02 | 54.32 | |
| AATAttack Setting=Black-box, Target RL Policy=PPO2026.02 | 54.32 | |
| AATAttack Setting=White-box, Target RL Policy=DQN2026.02 | 50.56 | |
| AATAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 50.56 | |
| TSGEAttack Setting=White-box, Target RL Policy=DQN2026.02 | 48.39 | |
| TSGEAttack Setting=Black-box, Target RL Policy=DQN2026.02 | 48.39 |