DISCOVER: Deep identification of symbolic open-form PDEs via enhanced reinforcement-learning