List your company

Offline-to-Online Reinforcement Learning on D4RL Antmaze (All Configurations)

74.51Success Rate (Large Diverse)ROAD
31.8742.9454.0165.08May 14, 2026

Evaluation results

MethodLinks
ROADBackbone=PEX2026.0574.5169.139289.6786.3599.83
Fixed Mixing RatioMixing Ratio=0.0, Backbone=PEX2026.0570.1564.3387.668795.1697.99
Decreasing Mixing StrategyBackbone=PEX2026.0569.9969.0193.2584.9951.7597
Fixed Mixing RatioMixing Ratio=0.5, Backbone=PEX2026.0569.8364.1886.4986.668.6897.34
Fixed Mixing RatioMixing Ratio=0.1, Backbone=PEX2026.0569.4866.6686.3388.8386.9999
Fixed Mixing RatioMixing Ratio=0.2, Backbone=PEX2026.0566.9768.3489.8286.8381.0197.66
Fixed Mixing RatioMixing Ratio=0.4, Backbone=PEX2026.0565.9867.0186.6785.496.1797.16
Uniform Mixing StrategyBackbone=PEX2026.0564.4764.7591.584.7542.2296.25
ROADMixing Ratio Strategy=ROAD, Algorithm=IQL2026.0563.1359.7583.6783.4972.1295.83
Fixed Mixing RatioMixing Ratio=0.3, Backbone=PEX2026.0562.5167.1886.8289.334398
UniformMixing Ratio Strategy=Uniform, Algorithm=IQL2026.0560.8346.3282.382.6725.6591.32
0.1Mixing Ratio Strategy=Fixed, Ratio=0.1, Algorithm=IQL2026.0560.3246.5178.6879.1769.6693.32
0.0Mixing Ratio Strategy=Fixed, Ratio=0.0, Algorithm=IQL2026.0556.9858.168082.563.4992.16
0.3Mixing Ratio Strategy=Fixed, Ratio=0.3, Algorithm=IQL2026.0556.0153.6681.6680.1638.593.34
0.2Mixing Ratio Strategy=Fixed, Ratio=0.2, Algorithm=IQL2026.0555.9853.1680.577.327.8393.16
DecreasingMixing Ratio Strategy=Decreasing, Algorithm=IQL2026.0552.1552.1781.8277.1648.8292.99
0.4Mixing Ratio Strategy=Fixed, Ratio=0.4, Algorithm=IQL2026.0550.6646.8282.678246.5194.33
0.5Mixing Ratio Strategy=Fixed, Ratio=0.5, Algorithm=IQL2026.0550.3555.1283.2978.6216.1394.37
BRMixing Ratio Strategy=BR, Algorithm=IQL2026.0546.8338.3481.8480.9980.6694.33
BRBackbone=PEX2026.0533.516488.589.2583.596.75