Neural Ordinary Differential Equations for Nonlinear System Identification
This work addresses system identification for dynamical systems, providing incremental improvements in accuracy and robustness over existing methods.
The paper tackles nonlinear system identification by comparing neural ordinary differential equations (NODE) with state-of-the-art methods, finding that NODEs improve prediction accuracy by an order of magnitude and are less sensitive to hyperparameters, though with a slight increase in inference computation.
Neural ordinary differential equations (NODE) have been recently proposed as a promising approach for nonlinear system identification tasks. In this work, we systematically compare their predictive performance with current state-of-the-art nonlinear and classical linear methods. In particular, we present a quantitative study comparing NODE's performance against neural state-space models and classical linear system identification methods. We evaluate the inference speed and prediction performance of each method on open-loop errors across eight different dynamical systems. The experiments show that NODEs can consistently improve the prediction accuracy by an order of magnitude compared to benchmark methods. Besides improved accuracy, we also observed that NODEs are less sensitive to hyperparameters compared to neural state-space models. On the other hand, these performance gains come with a slight increase of computation at the inference time.