Cartpole is often used in reinforcement learning research: it's easy to simulate and fast to run, but unlike pattern recognition tasks, it requires constant, fine-grained adjustments rather than a ...