List your company

Offline Reinforcement Learning on D4RL Medium-Replay Hopper

110.6Normalized ScoreNEUBAY
34.6854.3974.193.81May 24, 2023Nov 23, 2023May 24, 2024Nov 23, 2024May 25, 2025Nov 24, 2025May 27, 2026

Evaluation results

MethodLinks
NEUBAYMethod Category=Ours2025.12110.6
PSPOAveraged across=4 random seeds2026.05110
SUMOMethod Category=Conservative model-based2025.12109.9
VIPOMethod Category=Conservative model-based2025.12109.6
RRPINumber of random seeds=42026.03106.6
PMDBNumber of random seeds=42026.03106.2
PMDBMethod Type=Model-based2023.10106.2
PMDB2026.05106.2
ADMPOMethod Category=Conservative model-based2025.12104.4
AMGNumber of random seeds=42026.03104.4
ADM2026.05104.4
CBOPMethod Category=Bayesian-inspired2025.12104.3
MoMoMethod Category=Conservative model-based2025.12104
LEQMethod Category=Conservative model-based2025.12103.9
MOBILEMethod Category=Conservative model-based2025.12103.9
MOPOMethod Category=Conservative model-based2025.12103.5
ATACMethod Type=Model-free2023.10102.5
DMGNumber of random seeds=42026.03101.9
DMG2026.05101.9
SPAR2026.05101.9
Count-MORLMethod Type=Model-based2023.10101.7
QIPO2026.05101.2
EDACMethod Category=Model-free2025.12101
EDACMethod Type=Model-free2023.10101
DDImplementation=Original2023.06100
Decision Diffuser2025.12100
DAC2026.0599.8
RAMBONumber of random seeds=42026.0399.5
RAMBO2026.0599.5
CROPMethod Type=Model-based2023.1099.4
APE-VMethod Category=Conservative model-based2025.1298.5
MPDiffuser+Rankranking=true2025.1298.3
MPDiffuserranking=false2025.1298.2
EPQNumber of random seeds=42026.0397.8
EPQ2026.0597.8
IQLAccess=True reward functions2024.0497.4
Diffuser2023.0696.8
Diffuser2025.1296.8
RAMBOMethod Type=Model-based2023.1096.6
DDImplementation=reproduced2023.0695.3
CQL2023.0695
CQLModel Category=Value-based2026.0295
DMMModel Category=SSM-based2026.0295
CQL2023.0695
IQL2023.0694.7
IQLMethod Type=Model-free2023.1094.7
MoDAPMethod Category=Bayesian-inspired2025.1294.2
MOREL2023.0693.6
RAMBOMethod Category=Model-free2025.1292.8
D-MPC2025.1292.5
Planner2025.1292.1
TT2023.0691.5
TT2023.0691.5
Decision Stacks2023.0689.5
COMBOMethod Category=Model-free2025.1289.5
COMBO2025.1289.5
COMBOMethod Type=Model-based2023.1089.5
EDTModel Category=Transformer-based2026.0289.4
DMambasource=Lv et al., 20242023.0689.1
OPPOAccess=Scripted preferences, Implementation=Reported2024.0488.9
DCModel Category=SSM-based2026.0287.8
TAP2023.0687.3
Decision Stacks2023.0686.9
CQLMethod Category=Model-free2025.1286.3
CQLNumber of random seeds=42026.0386.3
CQL2026.0586.3
MambaDM2023.0686.1
MAPLEMethod Category=Bayesian-inspired2025.1285
PTOracle status=false, Implementation source=Kim et al. [28]2023.0584.54
TT2023.0684.4
RATE2023.0683.7
IQLOracle status=true, Implementation source=Kim et al. [28]2023.0583.06
DT2023.0682.7
DT2025.1282.7
DT2023.0682.7
DMambasource=Ota, 20242023.0682.6
DTModel Category=Transformer-based2026.0282.5
IDQL2025.1282.4
PTAccess=Scripted preferences2024.0482
MOReLNumber of random seeds=42026.0381.8
MOReL2026.0581.8
DMModel Category=SSM-based2026.0281.7
Diffuser2023.0680.9
DT2023.0678.3
IPLOracle status=false, Implementation source=Ours2023.0573.57
MROracle status=false, Implementation source=reimpl.2023.0570.2
FQL2026.0564.8
BREXOracle status=false, Implementation source=reimpl.2023.0562
TD3+BCModel Category=Value-based2026.0260.9
LSTMOracle status=false, Implementation source=Kim et al. [28]2023.0557.88
CAMPAccess=Scripted preferences2024.0456.2
IGDF2025.1254.9
DARA2025.1253.5
QLDTModel Category=Transformer-based2026.0252.1
DROCO2025.1251.6
IQL*variant=tuned/reproduced2025.1251.3
DS4Model Category=SSM-based2026.0249.6
OPRLAccess=Scripted preferences2024.0445.6
BOSA2025.1240.6
CQL*variant=tuned/reproduced2025.1237.6