Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sabug Jr., Lorenzo, Kerrigan, Eric
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2512.13217
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911319492395008
author Sabug Jr., Lorenzo
Kerrigan, Eric
author_facet Sabug Jr., Lorenzo
Kerrigan, Eric
contents We revisit the problem of physics-informed regression, and propose a method that directly computes the state at the prediction point, simultaneously with the derivative and curvature information of the existing samples. We frame each prediction as a constrained optimisation problem, leveraging multivariate Taylor series expansions and explicitly enforcing physical laws. Each individual query can be processed with low computational cost without any pre- or re-training, in contrast to global function approximator-based solutions such as neural networks. Our comparative benchmarks on a reaction-diffusion system show competitive predictive accuracy relative to a neural network-based solution, while completely eliminating the need for long training loops, and remaining robust to changes in the sampling layout.
format Preprint
id arxiv_https___arxiv_org_abs_2512_13217
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Rethinking Physics-Informed Regression Beyond Training Loops and Bespoke Architectures
Sabug Jr., Lorenzo
Kerrigan, Eric
Optimization and Control
Machine Learning
62J02, 62G08
We revisit the problem of physics-informed regression, and propose a method that directly computes the state at the prediction point, simultaneously with the derivative and curvature information of the existing samples. We frame each prediction as a constrained optimisation problem, leveraging multivariate Taylor series expansions and explicitly enforcing physical laws. Each individual query can be processed with low computational cost without any pre- or re-training, in contrast to global function approximator-based solutions such as neural networks. Our comparative benchmarks on a reaction-diffusion system show competitive predictive accuracy relative to a neural network-based solution, while completely eliminating the need for long training loops, and remaining robust to changes in the sampling layout.
title Rethinking Physics-Informed Regression Beyond Training Loops and Bespoke Architectures
topic Optimization and Control
Machine Learning
62J02, 62G08
url https://arxiv.org/abs/2512.13217