EUVIMEDEUVIMED entdecken
Crossref

Differentiable Programming for Differential Equations: A Review

Facundo Sapienza, Jordi Bolibar, Frank Schäfer, Brian Groenke, Avik Pal, Victor Boussange, Patrick Heimbach, Giles Hooker, Fernando Pérez, Per-Olof Persson, Christopher Rackauckas

SIAM Review · 31.12.2026 · 68 (4) · 749-819

DOI: 10.1137/24m1669062 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SIAM Review, online erschienen am 12.11.2026. Verfasst von Facundo Sapienza und 10 weiteren Autor:innen. Die Arbeit nennt 342 Literaturangaben. Gefördert von: Helmholtz Einstein International Berlin Research School in Data Science, Air Force Research Laboratory, Defense Advanced Research Projects Agency.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The differentiable programming paradigm is a cornerstone of modern scientific computing. It refers to numerical methods for computing the gradient of a numerical model’s output. Many scientific models are based on differential equations, in which differentiable programming plays a crucial role in calculating model sensitivities, inverting model parameters, and training hybrid models that combine differential equations with data-driven approaches. Furthermore, recognizing the strong synergies between inverse methods and machine learning offers the opportunity to establish a coherent framework applicable to both fields. Differentiating functions based on the numerical solution of differential equations is nontrivial. Numerous methods based on a wide variety of paradigms have been proposed in the literature, each with pros and cons specific to the type of problem investigated. Here, we provide a comprehensive review of existing techniques to compute derivatives of numerical solutions of differential equations. We first discuss the importance of gradients of solutions of differential equations in a variety of scientific domains. Second, we lay out and compare the mathematical foundations of the various approaches. Third, we cover computational considerations and explore the solutions available in modern scientific software. Last but not least, we provide best practices and recommendations for practitioners. We hope that this work accelerates the fusion of scientific models and data and fosters a modern approach to scientific modeling.

Förderung

  • Helmholtz Einstein International Berlin Research School in Data Science
  • Air Force Research Laboratory
  • Defense Advanced Research Projects Agency
  • Office of Naval Research
  • National Nuclear Security Administration
  • Stichting voor de Technische Wetenschappen
  • National Science Foundation
  • Nederlandse Organisatie voor Wetenschappelijk Onderzoek

Zitieren

APA

Sapienza, F., Bolibar, J., Schäfer, F., Groenke, B., Pal, A., Boussange, V., Heimbach, P., Hooker, G., Pérez, F., Persson, P., & Rackauckas, C. (2026). Differentiable Programming for Differential Equations: A Review. SIAM Review, 68(4), 749-819. https://doi.org/10.1137/24m1669062

Vancouver

Sapienza F, Bolibar J, Schäfer F, Groenke B, Pal A, Boussange V, et al. Differentiable Programming for Differential Equations: A Review. SIAM Review. 2026;68(4):749-819. doi:10.1137/24m1669062

BibTeX
@article{Sapienza2026,
  author = {Sapienza, Facundo and Bolibar, Jordi and Schäfer, Frank and Groenke, Brian and Pal, Avik and Boussange, Victor and Heimbach, Patrick and Hooker, Giles and Pérez, Fernando and Persson, Per-Olof and Rackauckas, Christopher},
  title = {Differentiable Programming for Differential Equations: A Review},
  journal = {SIAM Review},
  year = {2026},
  volume = {68},
  number = {4},
  pages = {749--819},
  doi = {10.1137/24m1669062}
}
Kostenlos starten →