Nonlinearity encoding to improve extrapolation capabilities for unobserved physical states

The fundamental goal of machine learning (ML) in physical science is to predict the physical properties of unobserved states. However, an accurate prediction for input data outside of training distributions is a challenging problem in ML due to the nonlinearities in input and target dynamics. For an...

Full description

Saved in:
Bibliographic Details
Published in:Physical chemistry chemical physics : PCCP 2022-01, Vol.24 (3), p.13-134
Main Authors: Na, Gyoung S, Jang, Seunghun, Chang, Hyunju
Format: Article
Language:English
Subjects:
Algorithms
Coders
Extrapolation
Machine learning
Many body problem
Material properties
Nonlinearity
Physical properties
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The fundamental goal of machine learning (ML) in physical science is to predict the physical properties of unobserved states. However, an accurate prediction for input data outside of training distributions is a challenging problem in ML due to the nonlinearities in input and target dynamics. For an accurate extrapolation of ML algorithms, we propose a new data-driven method that encodes the nonlinearities of physical systems into input representations. Based on the proposed encoder, a given physical system is described as linear-like functions that are easy to extrapolate. By applying the proposed encoder, the extrapolation errors were significantly reduced by 48.39% and 40.04% in n -body problem and materials property prediction, respectively. Performance improvement by nonlinearity encoding in an extrapolation task.
ISSN:1463-9076
1463-9084
DOI:10.1039/d1cp04450h