ResearchBenchmarksContinuous Control on reacher3da RandomTargetFollow0.467Average RewardDDPG-0.066520.071990.21050.34901Sep 9, 2015Evaluation ResultsMethodMethodLinksAverage RewardBest RewardDDPGObservation Space=low-...Observation Space=low-dimensional2015.090.4670.739DDPGObservation Space=pixelsObservation Space=pixels2015.090.3740.735DPGSetup=original with re...Setup=original with replay buffer and batch normalization2015.09-0.0460.158