List your company
ResearchFormerly WizWand

Offline Reinforcement Learning on Hopper

2,116.2Average ReturnUWMSG
-80.592489.7291,060.051,630.371Oct 23, 2023
Updated 8mo ago

Evaluation results

MethodLinks
UWMSGAttack Type=Random Dynamics2023.102,116.2
UWMSGAttack Type=Random Reward2023.102,021.1
MSGAttack Type=Random Reward2023.101,599.6
MSGAttack Type=Random Dynamics2023.101,552.7
UWMSGAttack Type=Adversarial Dynamics2023.10931.2
UWMSGAttack Type=Adversarial Reward2023.10751.4
MSGAttack Type=Adversarial Dynamics2023.10717.7
MSGAttack Type=Adversarial Reward2023.10651
SAC-NAttack Type=Random Reward2023.10178.8
SAC-NAttack Type=Adversarial Reward2023.10111.8
EDACAttack Type=Random Reward2023.10107.7
EDACAttack Type=Adversarial Reward2023.1029.2
EDACAttack Type=Random Dynamics2023.105.9
EDACAttack Type=Adversarial Dynamics2023.105.9
SAC-NAttack Type=Random Dynamics2023.103.9
SAC-NAttack Type=Adversarial Dynamics2023.103.9