{"title":"Machine Learning-Based Direct Solver for One-To-Many Problems on Temporal Shaping of Electron Beams","authors":"Jinyu Wan, Y. Jiao, Juhao Wu","doi":"10.21203/RS.3.RS-524222/V1","DOIUrl":null,"url":null,"abstract":"\n To control the temporal profile of an electron beam to meet requirements of various advanced scientific applications, a widely-used technique is to manipulate the dispersion terms which turns out to be one-to-many problems. Due to their intrinsic one-to-many property, current popular stochastic optimization approaches on temporal shaping are not very effective, for being trapped into local optima or suggesting only one solution. Here we propose a real-time solver for one-to-many problems of temporal shaping, with the aid of a semi-supervised machine learning method, the conditional generative adversarial network (CGAN). We demonstrate that the CGAN solver can learn the one-to-many dynamics and is able to accurately and quickly predict the required dispersion terms for different custom temporal profiles. This machine learning-based solver overcomes the limitation of the stochastic optimization methods and is expected to have the potential for wide applications to one-to-many problems in other scientific fields.","PeriodicalId":8436,"journal":{"name":"arXiv: Accelerator Physics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2021-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv: Accelerator Physics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.21203/RS.3.RS-524222/V1","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
To control the temporal profile of an electron beam to meet requirements of various advanced scientific applications, a widely-used technique is to manipulate the dispersion terms which turns out to be one-to-many problems. Due to their intrinsic one-to-many property, current popular stochastic optimization approaches on temporal shaping are not very effective, for being trapped into local optima or suggesting only one solution. Here we propose a real-time solver for one-to-many problems of temporal shaping, with the aid of a semi-supervised machine learning method, the conditional generative adversarial network (CGAN). We demonstrate that the CGAN solver can learn the one-to-many dynamics and is able to accurately and quickly predict the required dispersion terms for different custom temporal profiles. This machine learning-based solver overcomes the limitation of the stochastic optimization methods and is expected to have the potential for wide applications to one-to-many problems in other scientific fields.