DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Parameter uncertainties for imperfect surrogate models in the low-noise regime

Journal Article · · Machine Learning: Science and Technology

Abstract Bayesian regression determines model parameters by minimizing the expected loss, an upper bound to the true generalization error. However, this loss ignores model form error, or misspecification, meaning parameter uncertainties are significantly underestimated and vanish in the large data limit. As misspecification is the main source of uncertainty for surrogate models of low-noise calculations, such as those arising in atomistic simulation, predictive uncertainties are systematically underestimated. We analyze the true generalization error of misspecified, near-deterministic surrogate models, a regime of broad relevance in science and engineering. We show that posterior parameter distributions must cover every training point to avoid a divergence in the generalization error and design a compatible ansatz which incurs minimal overhead for linear models. The approach is demonstrated on model problems before application to thousand-dimensional datasets in atomistic machine learning. Our efficient misspecification-aware scheme gives accurate prediction and bounding of test errors in terms of parameter uncertainties, allowing this important source of uncertainty to be incorporated in multi-scale computational workflows.

Sponsoring Organization:
USDOE
OSTI ID:
2499832
Journal Information:
Machine Learning: Science and Technology, Journal Name: Machine Learning: Science and Technology Journal Issue: 1 Vol. 6; ISSN 2632-2153
Publisher:
IOP PublishingCopyright Statement
Country of Publication:
United Kingdom
Language:
English

References (30)

Machine Learning Interatomic Potentials as Emerging Tools for Materials Science journal September 2019
Managing computational complexity using surrogate models: a critical review journal April 2020
Recent advances and applications of surrogate models for finite element method computations: a review journal July 2022
Uncertainty Quantification in Atomistic Modeling of Metals and Its Effect on Mesoscale and Continuum Modeling: A Review journal October 2020
Active learning of linearly parametrized interatomic potentials journal December 2017
From CP-FFT to CP-RNN: Recurrent neural network surrogate model of crystal plasticity journal November 2022
Uncertainty quantification in scientific machine learning: Methods, metrics, and comparisons journal March 2023
Scalable Bayesian Uncertainty Quantification for Neural Network Potentials: Promise and Pitfalls journal April 2023
Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17 journal November 2012
Reinforcing materials modelling by encoding the structures of defects in crystalline solids into distortion scores journal September 2020
Machine-learned multi-system surrogate models for materials prediction journal April 2019
Uncertainty quantification in molecular simulations with dropout neural network potentials journal August 2020
On-the-fly active learning of interpretable Bayesian force fields for atomistic rare events journal March 2020
Complex strengthening mechanisms in the NbMoTaW multi-principal element alloy journal June 2020
Performant implementation of the atomic cluster expansion (PACE) and application to copper and silicon journal June 2021
Training data selection for accuracy and transferability of interatomic potentials journal September 2022
Quantum chemistry structures and properties of 134 kilo molecules journal August 2014
Linear graphlet models for accurate and interpretable cheminformatics journal January 2024
Extending the accuracy of the SNAP interatomic potential form journal June 2018
An entropy-maximization approach to automated training set generation for interatomic potentials journal September 2020
Uncertainty estimation for molecular dynamics and sampling journal February 2021
Uncertainty quantification in atomistic simulations of silicon using interatomic potentials journal August 2024
Uncertainty quantification by direct propagation of shallow ensembles journal July 2024
Uncertainty quantification in classical molecular dynamics
  • Wan, Shunzhou; Sinclair, Robert C.; Coveney, Peter V.
  • Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 379, Issue 2197 https://doi.org/10.1098/rsta.2020.0082
journal March 2021
Self-Consistent Equations Including Exchange and Correlation Effects journal November 1965
Exploring the robust extrapolation of high-dimensional machine learning potentials journal April 2022
Molecular Dynamics with On-the-Fly Machine Learning of Quantum-Mechanical Forces journal March 2015
Machine learning surrogate models for prediction of point defect vibrational entropy journal June 2020
Efficient and transferable machine learning potentials for the simulation of crystal defects in bcc Fe and W journal October 2021
Bayesian inference in physics journal September 2011

Related Subjects