List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on Walker2d

2,278.9Average ReturnUWMSG
-96.46520.221,136.91,753.58Oct 23, 2023
Updated 8mo ago

Evaluation results

MethodLinks
UWMSGAttack Type=Random Dynamics2023.102,278.9
UWMSGAttack Type=Random Reward2023.102,189.7
MSGAttack Type=Random Dynamics2023.102,122.8
UWMSGAttack Type=Adversarial Reward2023.101,433.4
UWMSGAttack Type=Adversarial Dynamics2023.10946
MSGAttack Type=Adversarial Reward2023.10605.8
MSGAttack Type=Random Reward2023.10539.1
MSGAttack Type=Adversarial Dynamics2023.10506
EDACAttack Type=Adversarial Reward2023.1061
SAC-NAttack Type=Adversarial Reward2023.109.5
SAC-NAttack Type=Random Dynamics2023.10-3.1
EDACAttack Type=Random Dynamics2023.10-3.6
EDACAttack Type=Random Reward2023.10-3.7
SAC-NAttack Type=Random Reward2023.10-3.8
EDACAttack Type=Adversarial Dynamics2023.10-4.8
SAC-NAttack Type=Adversarial Dynamics2023.10-5.1