Benchmarking Energy-Conserving Neural Networks for Learning Dynamics from Data
arXiv:2012.02334
Abstract
The last few years have witnessed an increased interest in incorporating physics-informed inductive bias in deep learning frameworks. In particular, a growing volume of literature has been exploring ways to enforce energy conservation while using neural networks for learning dynamics from observed time-series data. In this work, we survey ten recently proposed energy-conserving neural network models, including HNN, LNN, DeLaN, SymODEN, CHNN, CLNN and their variants. We provide a compact derivation of the theory behind these models and explain their similarities and differences. Their performance are compared in 4 physical systems. We point out the possibility of leveraging some of these energy-conserving models to design energy-based controllers.
References in corpus (10)
- Decoupled Weight Decay Regularization
- Hamiltonian Neural Networks
- Why Deep Neural Networks for Function Approximation?
- Symplectic ODE-Net: Learning Hamiltonian Dynamics with Control
- Hamiltonian Graph Networks with ODE Integrators
- Symplectic Neural Networks in Taylor Series Form for Hamiltonian Systems
- Dissipative SymODEN: Encoding Hamiltonian Dynamics with Dissipation and Control into Deep Learning
- Deep Hamiltonian networks based on symplectic integrators
- LagNetViP: A Lagrangian Neural Network for Video Prediction
- Lectures on Constrained Systems