000607614 001__ 607614
000607614 005__ 20250115142155.0
000607614 0247_ $$2INSPIRETeX$$aKaiser:2024hpa
000607614 0247_ $$2inspire$$ainspire:2745736
000607614 0247_ $$2arXiv$$aarXiv:2401.05815
000607614 0247_ $$2datacite_doi$$a10.3204/PUBDB-2024-01950
000607614 037__ $$aPUBDB-2024-01950
000607614 041__ $$aEnglish
000607614 082__ $$a530
000607614 088__ $$2arXiv$$aarXiv:2401.05815
000607614 1001_ $$0P:(DE-H253)PIP1095111$$aKaiser, Jan$$b0$$eCorresponding author$$udesy
000607614 245__ $$aBridging the gap between machine learning and particle accelerator physics with high-speed, differentiable simulations
000607614 260__ $$c2024
000607614 3367_ $$0PUB:(DE-HGF)25$$2PUB:(DE-HGF)$$aPreprint$$bpreprint$$mpreprint$$s1717145644_860015
000607614 3367_ $$2ORCID$$aWORKING_PAPER
000607614 3367_ $$028$$2EndNote$$aElectronic Article
000607614 3367_ $$2DRIVER$$apreprint
000607614 3367_ $$2BibTeX$$aARTICLE
000607614 3367_ $$2DataCite$$aOutput Types/Working Paper
000607614 500__ $$aPhys. Rev. Accel. Beams 27 (2024) 054601. 16 pages, 9 figures, 3 tables
000607614 520__ $$aMachine learning has emerged as a powerful solution to the modern challenges in accelerator physics. However, the limited availability of beam time, the computational cost of simulations, and the high dimensionality of optimization problems pose significant challenges in generating the required data for training state-of-the-art machine learning models. In this work, we introduce cheetah, a pytorch-based high-speed differentiable linear beam dynamics code. cheetah enables the fast collection of large datasets by reducing computation times by multiple orders of magnitude and facilitates efficient gradient-based optimization for accelerator tuning and system identification. This positions cheetah as a user-friendly, readily extensible tool that integrates seamlessly with widely adopted machine learning tools. We showcase the utility of cheetah through five examples, including reinforcement learning training, gradient-based beamline tuning, gradient-based system identification, physics-informed Bayesian optimization priors, and modular neural network surrogate modeling of space charge effects. The use of such a high-speed differentiable simulation code will simplify the development of machine learning-based methods for particle accelerators and fast-track their integration into everyday operations of accelerator facilities.
000607614 536__ $$0G:(DE-HGF)POF4-621$$a621 - Accelerator Research and Development (POF4-621)$$cPOF4-621$$fPOF IV$$x0
000607614 536__ $$0G:(DE-HGF)2020_InternLabs-0011$$aInternLabs-0011 - HIR3X - Helmholtz International Laboratory on Reliability, Repetition, Results at the most advanced X-ray Sources (2020_InternLabs-0011)$$c2020_InternLabs-0011$$x1
000607614 536__ $$0G:(DE-HGF)2020_ZT-I-PF-5-6$$aZT-I-PF-5-6 - Autonomous Accelerator (AA) (2020_ZT-I-PF-5-6)$$c2020_ZT-I-PF-5-6$$x2
000607614 588__ $$aDataset connected to CrossRef, INSPIRE, Journals: bib-pubdb1.desy.de
000607614 650_7 $$2INSPIRE$$aspace charge
000607614 650_7 $$2INSPIRE$$aaccelerator
000607614 650_7 $$2INSPIRE$$amachine learning
000607614 650_7 $$2INSPIRE$$aoptimization
000607614 650_7 $$2INSPIRE$$areinforcement learning
000607614 650_7 $$2INSPIRE$$amodular
000607614 650_7 $$2INSPIRE$$aBayesian
000607614 650_7 $$2INSPIRE$$aneural network
000607614 693__ $$0EXP:(DE-H253)ARES-20200101$$1EXP:(DE-H253)SINBAD-20200101$$5EXP:(DE-H253)ARES-20200101$$aSINBAD$$eAccelerator Research Experiment at SINBAD$$x0
000607614 7001_ $$aXu, Chenran$$b1
000607614 7001_ $$0P:(DE-H253)PIP1087213$$aEichler, Annika$$b2
000607614 7001_ $$aSantamaria Garcia, Andrea$$b3
000607614 8564_ $$uhttps://bib-pubdb1.desy.de/record/607614/files/2401.05815v1.pdf$$yOpenAccess
000607614 8564_ $$uhttps://bib-pubdb1.desy.de/record/607614/files/2401.05815v1.pdf?subformat=pdfa$$xpdfa$$yOpenAccess
000607614 909CO $$ooai:bib-pubdb1.desy.de:607614$$pdnbdelivery$$pdriver$$pVDB$$popen_access$$popenaire
000607614 9101_ $$0I:(DE-588b)2008985-5$$6P:(DE-H253)PIP1095111$$aDeutsches Elektronen-Synchrotron$$b0$$kDESY
000607614 9101_ $$0I:(DE-588b)2008985-5$$6P:(DE-H253)PIP1087213$$aDeutsches Elektronen-Synchrotron$$b2$$kDESY
000607614 9101_ $$0I:(DE-588)1043621512$$6P:(DE-H253)PIP1087213$$aEuropean XFEL$$b2$$kXFEL.EU
000607614 9131_ $$0G:(DE-HGF)POF4-621$$1G:(DE-HGF)POF4-620$$2G:(DE-HGF)POF4-600$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$aDE-HGF$$bForschungsbereich Materie$$lMaterie und Technologie$$vAccelerator Research and Development$$x0
000607614 9141_ $$y2024
000607614 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
000607614 915__ $$0LIC:(DE-HGF)CCBY4$$2HGFVOC$$aCreative Commons Attribution CC BY 4.0
000607614 915__ $$0StatID:(DE-HGF)0580$$2StatID$$aPublished
000607614 9201_ $$0I:(DE-H253)MSK-20120731$$kMSK$$lStrahlkontrollen$$x0
000607614 980__ $$apreprint
000607614 980__ $$aVDB
000607614 980__ $$aUNRESTRICTED
000607614 980__ $$aI:(DE-H253)MSK-20120731
000607614 9801_ $$aFullTexts