KIT | KIT-Bibliothek | Impressum | Datenschutz

Bridging the gap between machine learning and particle accelerator physics with high-speed, differentiable simulations

Kaiser, Jan; Xu, Chenran ORCID iD icon 1; Eichler, Annika; Santamaria Garcia, Andrea ORCID iD icon 2
1 Institut für Beschleunigerphysik und Technologie (IBPT), Karlsruher Institut für Technologie (KIT)
2 Laboratorium für Applikationen der Synchrotronstrahlung (LAS), Karlsruher Institut für Technologie (KIT)

Abstract:

Machine learning has emerged as a powerful solution to the modern challenges in accelerator physics. However, the limited availability of beam time, the computational cost of simulations, and the high dimensionality of optimization problems pose significant challenges in generating the required data for training state-of-the-art machine learning models. In this work, we introduce cheetah, a pytorch-based high-speed differentiable linear beam dynamics code. cheetah enables the fast collection of large datasets by reducing computation times by multiple orders of magnitude and facilitates efficient gradient-based optimization for accelerator tuning and system identification. This positions cheetah as a user-friendly, readily extensible tool that integrates seamlessly with widely adopted machine learning tools. We showcase the utility of cheetah through five examples, including reinforcement learning training, gradient-based beamline tuning, gradient-based system identification, physics-informed Bayesian optimization priors, and modular neural network surrogate modeling of space charge effects. The use of such a high-speed differentiable simulation code will simplify the development of machine learning-based methods for particle accelerators and fast-track their integration into everyday operations of accelerator facilities.


Verlagsausgabe §
DOI: 10.5445/IR/1000171541
Veröffentlicht am 13.06.2024
Originalveröffentlichung
DOI: 10.1103/PhysRevAccelBeams.27.054601
Scopus
Zitationen: 4
Web of Science
Zitationen: 4
Dimensions
Zitationen: 4
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Beschleunigerphysik und Technologie (IBPT)
Laboratorium für Applikationen der Synchrotronstrahlung (LAS)
Publikationstyp Zeitschriftenaufsatz
Publikationsmonat/-jahr 05.2024
Sprache Englisch
Identifikator ISSN: 2469-9888
KITopen-ID: 1000171541
HGF-Programm 54.11.11 (POF IV, LK 01) Accelerator Operation, Research and Development
Erschienen in Physical Review Accelerators and Beams
Verlag American Physical Society (APS)
Band 27
Heft 5
Seiten Art.-Nr.: 054601
Projektinformation Autonomous Accelerator (HGF, HGF IVF2016 STRATZUK, ZT-I-PF-5-6)
Vorab online veröffentlicht am 28.05.2024
Nachgewiesen in Dimensions
Scopus
Web of Science
Relationen in KITopen
Globale Ziele für nachhaltige Entwicklung Ziel 4 – Hochwertige BildungZiel 9 – Industrie, Innovation und Infrastruktur
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page