List your company

Control Task on OpenAI Gym BipedalWalker (Statistical Comparison)

-2.0642T-statisticPOEM
-2.16741-2.115805-2.0642-2.012595Jan 21, 2026

Evaluation results

MethodLinks
POEMtraining timesteps=1,500,000, experiments=10, evaluation episodes=15, baseline=PPO2026.01-2.06420.0495—