Differentiating Through Linear Solvers

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hovland, Paul, Hückelheim, Jan
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916237614776320
author Hovland, Paul
Hückelheim, Jan
author_facet Hovland, Paul
Hückelheim, Jan
contents Computer programs containing calls to linear solvers are a known challenge for automatic differentiation. Previous publications advise against differentiating through the low-level solver implementation, and instead advocate for high-level approaches that express the derivative in terms of a modified linear system that can be solved with a separate solver call. Despite this ubiquitous advice, we are not aware of prior work comparing the accuracy of both approaches. With this article we thus empirically study a simple question: What happens if we ignore common wisdom, and differentiate through linear solvers?
format Preprint
id arxiv_https___arxiv_org_abs_2404_17039
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Differentiating Through Linear Solvers
Hovland, Paul
Hückelheim, Jan
Mathematical Software
Numerical Analysis
Computer programs containing calls to linear solvers are a known challenge for automatic differentiation. Previous publications advise against differentiating through the low-level solver implementation, and instead advocate for high-level approaches that express the derivative in terms of a modified linear system that can be solved with a separate solver call. Despite this ubiquitous advice, we are not aware of prior work comparing the accuracy of both approaches. With this article we thus empirically study a simple question: What happens if we ignore common wisdom, and differentiate through linear solvers?
title Differentiating Through Linear Solvers
topic Mathematical Software
Numerical Analysis
url https://arxiv.org/abs/2404.17039