Weak-Form Latent Space Dynamics Identification

RESOURCE

Abstract

This software showcases the enhanced capabilities of the Latent Space Dynamics Identification (LaSDI) algorithm through the application of the weak form, resulting in WLaSDI. WLaSDI first compresses the data, then projects it onto test functions, and subsequently learns the local latent space models. Notably, WLaSDI demonstrates significantly improved robustness to noise. Using weak-form equation learning techniques, WLaSDI achieves local latent space modeling. Compared to the standard sparse identification of nonlinear dynamics (SINDy) used in LaSDI, the variance reduction of the weak form ensures robust and precise latent space recovery, enabling fast, robust, and accurate simulations. We demonstrate the efficacy of WLaSDI against LaSDI using several common benchmark examples, including viscid and inviscid Burgers', radial advection, and heat conduction. For instance, in 1D inviscid Burgers' simulations with up to 100% Gaussian white noise, WLaSDI maintains relative errors consistently below 6%, whereas LaSDI errors can exceed 10,000%. Similarly, in radial advection simulations, WLaSDI keeps relative errors below 16%, compared to potential errors of up to 10,000% with LaSDI. Additionally, WLaSDI achieves significant speedups, such as a 140X speedup in 1D Burgers' simulations compared to the corresponding full order model.
Developers:
Choi, Youngsoo [1] He, Xiaolong [1] Tran, Chi [1] Bortz, David [2]
  1. Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
  2. University of Colorado, Boulder
Release Date:
2024-07-05
Project Type:
Open Source, Publicly Available Repository
Software Type:
Scientific
Version:
0.1
Licenses:
MIT License
Sponsoring Org.:
Code ID:
142351
Site Accession Number:
LLNL-CODE-867254
Research Org.:
Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
Country of Origin:
United States

RESOURCE

Citation Formats

Choi, Youngsoo, He, Xiaolong, Tran, Chi K., and Bortz, David. Weak-Form Latent Space Dynamics Identification. Computer Software. https://github.com/MathBioCU/PyWLaSDI. USDOE National Nuclear Security Administration (NNSA). 05 Jul. 2024. Web. doi:10.11578/dc.20240904.1.
Choi, Youngsoo, He, Xiaolong, Tran, Chi K., & Bortz, David. (2024, July 05). Weak-Form Latent Space Dynamics Identification. [Computer software]. https://github.com/MathBioCU/PyWLaSDI. https://doi.org/10.11578/dc.20240904.1.
Choi, Youngsoo, He, Xiaolong, Tran, Chi K., and Bortz, David. "Weak-Form Latent Space Dynamics Identification." Computer software. July 05, 2024. https://github.com/MathBioCU/PyWLaSDI. https://doi.org/10.11578/dc.20240904.1.
@misc{ doecode_142351,
title = {Weak-Form Latent Space Dynamics Identification},
author = {Choi, Youngsoo and He, Xiaolong and Tran, Chi K. and Bortz, David},
abstractNote = {This software showcases the enhanced capabilities of the Latent Space Dynamics Identification (LaSDI) algorithm through the application of the weak form, resulting in WLaSDI. WLaSDI first compresses the data, then projects it onto test functions, and subsequently learns the local latent space models. Notably, WLaSDI demonstrates significantly improved robustness to noise. Using weak-form equation learning techniques, WLaSDI achieves local latent space modeling. Compared to the standard sparse identification of nonlinear dynamics (SINDy) used in LaSDI, the variance reduction of the weak form ensures robust and precise latent space recovery, enabling fast, robust, and accurate simulations. We demonstrate the efficacy of WLaSDI against LaSDI using several common benchmark examples, including viscid and inviscid Burgers', radial advection, and heat conduction. For instance, in 1D inviscid Burgers' simulations with up to 100% Gaussian white noise, WLaSDI maintains relative errors consistently below 6%, whereas LaSDI errors can exceed 10,000%. Similarly, in radial advection simulations, WLaSDI keeps relative errors below 16%, compared to potential errors of up to 10,000% with LaSDI. Additionally, WLaSDI achieves significant speedups, such as a 140X speedup in 1D Burgers' simulations compared to the corresponding full order model.},
doi = {10.11578/dc.20240904.1},
url = {https://doi.org/10.11578/dc.20240904.1},
howpublished = {[Computer Software] \url{https://doi.org/10.11578/dc.20240904.1}},
year = {2024},
month = {jul}
}