Discovering Interpretable Ordinary Differential Equations from Noisy Data