Inverse modified differential equations for discovery of dynamics
arXiv:2009.01058
Abstract
The combination of numerical integration and deep learning, i.e., ODE-net, has been successfully employed in a variety of applications. In this work, we introduce inverse modified differential equations (IMDE) to contribute to the behaviour and error analysis of discovery of dynamics using ODE-net. It is shown that the difference between the learned ODE and the truncated IMDE is bounded by the sum of learning loss and a discrepancy which can be made sub exponentially small. In addition, we deduce that the total error of ODE-net is bounded by the sum of discrete error and learning loss. Furthermore, with the help of IMDE, theoretical results on learning Hamiltonian system are derived. Several experiments are performed to numerically verify our theoretical results.
References in corpus (9)
- ANODE: Unconditionally Accurate Memory-Efficient Gradients for Neural ODEs
- Symplectic Neural Networks in Taylor Series Form for Hamiltonian Systems
- OnsagerNet: Learning Stable and Interpretable Dynamics using a Generalized Onsager Principle
- Continuous-in-Depth Neural Networks
- Deep Hamiltonian networks based on symplectic integrators
- Time Dependence in Non-Autonomous Neural ODEs
- ResNet After All? Neural ODEs and Their Numerical Solution
- Towards Understanding Normalization in Neural ODEs
- The Discovery of Dynamics via Linear Multistep Methods and Deep Learning: Error Estimation