List your company

Offline Reinforcement Learning on D4RL Adroit hammer-cloned

2,280Normalized ScoreEPQ
-90.888524.6311,140.151,755.669Nov 15, 2025Dec 1, 2025Dec 17, 2025Jan 2, 2026Jan 18, 2026Feb 3, 2026Feb 20, 2026

Evaluation results

MethodLinks
EPQPolicy Type=Gaussian, Seeds=82026.022,280
NEUBAYModel paradigm=Ours, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.121,440
FACPolicy Type=Flow, Seeds=82026.021,110
FQLPolicy Type=Flow, Seeds=82026.021,100
ReBRACPolicy Type=Gaussian, Seeds=82026.02670
ReBRACModel paradigm=Model-free, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.12500
VIPOModel paradigm=Conservative model-based, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.12210
CQLPolicy Type=Gaussian, Seeds=82026.02210
IDQLPolicy Type=Diffusion, Seeds=82026.02200
SRPOPolicy Type=Diffusion, Seeds=82026.02200
IQLModel paradigm=Model-free, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.12150
MOBILEModel paradigm=Conservative model-based, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.12150
MCQPolicy Type=Gaussian, Seeds=82026.02140
IQLPolicy Type=Gaussian, Seeds=82026.02110
CACPolicy Type=Diffusion, Seeds=82026.02100
TD3+BCPolicy Type=Gaussian, Seeds=82026.0280
ARMORModel paradigm=Conservative model-based, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.1270
MoMoModel paradigm=Conservative model-based, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.1270
BCModel paradigm=Model-free, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.1260
MOPOModel paradigm=Conservative model-based, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.1250
EDACModel paradigm=Model-free, pi_D=0.7, Number of seeds=6, Evaluation episodes=20, Evaluation step=final step2025.1230
IQLDomain=Adroit, Hyperparameter Tuning=individually tuned2025.1111.6
MXQLDomain=Adroit, Hyperparameter Tuning=individually tuned2025.1111.1
SAC-RNDPolicy Type=Gaussian, Seeds=82026.0210
BCDomain=Adroit, Hyperparameter Tuning=individually tuned2025.118.9
QQLDomain=Adroit, Hyperparameter Tuning=consistent2025.118
XQLDomain=Adroit, Hyperparameter Tuning=individually tuned2025.111.1
TD3+BCDomain=Adroit, Hyperparameter Tuning=individually tuned2025.110.9
CQLDomain=Adroit, Hyperparameter Tuning=individually tuned2025.110.3