pnnl/neural_ODE_ICLR2020
- Pacific Northwest National Laboratory
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
We show how to model discrete ordinary differential equations (ODE) with algebraic nonlinearities as deep neural networks with varying degrees of prior knowledge. We derive the stability guarantees of the network layers based on the implicit constraints imposed on the weight's eigenvalues. Moreover, we show how to use barrier methods to generically handle additional inequality constraints. We demonstrate the prediction accuracy of learned neural ODEs evaluated on open-loop simulations compared to ground truth dynamics with bi-linear terms.
- Short Name / Acronym:
- neural_ODE_ICLR2020
- Site Accession Number:
- Battelle IPID 31922-E
- Software Type:
- Scientific
- License(s):
- BSD 3-clause "New" or "Revised" License
- Research Organization:
- Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOEPrimary Award/Contract Number:AC05-76RL01830
- DOE Contract Number:
- AC05-76RL01830
- Code ID:
- 67085
- OSTI ID:
- code-67085
- Country of Origin:
- United States
Similar Records
scikit-SUNDAE ((SUN)DIALS Differential Algebraic Equations) [SWR-24-137]
IDA Version 2.2.0
pnnl/neuromancer
Software
·
Wed Oct 30 20:00:00 EDT 2024
·
OSTI ID:code-146982
IDA Version 2.2.0
Software
·
Wed Dec 08 00:00:00 EST 2004
·
OSTI ID:1245758
pnnl/neuromancer
Software
·
Tue Sep 21 20:00:00 EDT 2021
·
OSTI ID:code-64176