Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Surrogate construction via weight parameterization of residual neural networks

Journal Article · · Computer Methods in Applied Mechanics and Engineering

Surrogate model development is a critical step for uncertainty quantification or other sample-intensive tasks for complex computational models. Here, in this work, we develop a multi-output surrogate form using a class of neural networks (NNs) that employ shortcut connections, namely Residual NNs (ResNets). ResNets are known to regularize the surrogate learning problem and improve the efficiency and accuracy of the resulting surrogate. Inspired by the continuous, Neural ODE analogy, we augment ResNets with weight parameterization strategy with respect to ResNet depth. Weight-parameterized ResNets regularize the NN surrogate learning problem and allow better generalization with a drastically reduced number of learnable parameters. We demonstrate that weight-parameterized ResNets are more accurate and efficient than conventional feed-forward multi-layer perceptron networks. We also compare various options for parameterization of the weights as functions of ResNet depth. We demonstrate the results on both synthetic examples and a large scale earth system model of interest.

Research Organization:
Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)
Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA); USDOE Laboratory Directed Research and Development (LDRD) Program
Grant/Contract Number:
NA0003525
OSTI ID:
2476624
Report Number(s):
SAND--2024-15148J
Journal Information:
Computer Methods in Applied Mechanics and Engineering, Journal Name: Computer Methods in Applied Mechanics and Engineering Journal Issue: Part A Vol. 433; ISSN 0045-7825
Publisher:
ElsevierCopyright Statement
Country of Publication:
United States
Language:
English

References (10)

Learning high-dimensional parametric maps via reduced basis adaptive residual networks journal December 2022
Polynomial chaos expansion for sensitivity analysis journal July 2009
Stiff neural ordinary differential equations journal September 2021
Stable architectures for deep neural networks journal December 2017
Deep Residual Learning for Image Recognition conference June 2016
Spectral Representation and Reduced Order Modeling of the Dynamics of Stochastic Reaction Networks via Adaptive Data Partitioning journal January 2010
Dimensionality Reduction for Complex Models via Bayesian Compressive Sensing journal January 2014
The role of Stiffness in Training and Generalization of Resnets journal January 2023
Benchmarking and parameter sensitivity of physiological and vegetation dynamics using the Functionally Assembled Terrestrial Ecosystem Simulator (FATES) at Barro Colorado Island, Panama journal January 2020
Efficient surrogate modeling methods for large-scale Earth system models based on machine-learning techniques journal January 2019

Similar Records

Analysis of Neural Networks as Random Dynamical Systems
Technical Report · Fri Sep 01 00:00:00 EDT 2023 · OSTI ID:2430214

Layer-Parallel Training of Deep Residual Neural Networks
Journal Article · Wed Feb 05 23:00:00 EST 2020 · SIAM Journal on Mathematics of Data Science · OSTI ID:1618082

On the influence of over-parameterization in manifold based surrogates and deep neural operators
Journal Article · Mon Feb 13 23:00:00 EST 2023 · Journal of Computational Physics · OSTI ID:2421330