home › event - improving state-action space exploration in reinforcement learning using geometric properties