List your company

Offline Reinforcement Learning on D4RL Gym

35.6Return (Hopper, Random)SPQR-SAC-Min
2.42411.03719.6528.263Nov 2, 2021Jul 27, 2022Apr 20, 2023Jan 12, 2024Oct 5, 2024Jun 29, 2025Mar 23, 2026

Evaluation results

MethodLinks
SPQR-SAC-Min2024.0135.689.6————100.2104.9112.524.698.487.8118.2115.2101.133.574.8112.811469.788.5112109.1—
SAC-Min2024.0131.384.5————100.3101.8110.121.787.978.7116.7116.794.62867.5105.2107.163.984.5110.3102.9—
EDAC2024.0125.385.2————101.6101110.716.692.587.1114.7115.199.828.465.9106.8106.361.384.6110.1105.4—
SAC-RND2026.0319.682.6————91.197.2109.8—92.789.4104.6104.5105.327.666.4102.6108.151.281.2109.8107.418.7
CQL2021.1110.8—35.444.44262.45829.5111779.221.198.7———————————
S4RL-(N)Augmentation=Normal2021.1110.8—52.348.851.47978.935.4113.524.993.630.3112.2———————————
S4RL-(Adv)Augmentation=Adversarial2021.1110.7—53.948.651.778.181.336.8117.925.193.135107.1———————————
KFC++2021.1110.7—49.259.158.979.894.249125.517.610846.1115.3———————————
KFC2021.1110.4—48.655.958.179.990.648.612119.1102.148114———————————
IQL2026.0310.172.9————65.289.685.5—80.775.4112.196.997.519.55095.592.742.175108.8104.411.3
ReBRAC2026.038.5681.53————102.0595.82105.7—84.7880.67111.81112.15100.1629.5965.87105.22100.2950.2182.19104.73107.4920.22
MPCwDWMpre-trained policy=ReBRAC2026.038.5685.33————103.38103.14107.42—88.9195.87115.76116.65105.7629.970.05106.14105.3559.8985.149105.06108.7720.22
TD3+BC2026.038.570.3————60.955.187.8—77.768110.411090.330.954.793.489.14575109.697.92
CQL-Min2024.01770.9————86.648.6111774.532.698.7121.698.9635.444.4104.862.446.282.07109.9104.85—
CQL2026.035.373.6————61.986.396.9—79.576.8109.1109.394.231.146.997.39545.376.9106.5101.95.1
BC2024.013.749.9————54.116.653.91.370.920.390.1108.768.82.243.291.84437.662.9107.719.9—