Reinforcement learning for control: Performance, stability, and deep approximators | Steady Practice | SteadyPractice