[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117843-en":3,"doc-seo-117843-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},117843,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","Beam Detection Based on Machine Learning Algorithms","Free electron laser (FEL) beam spots on screens are determined precisely through a cascade of machine learning models. A self-constructed convolutional neural network performs transfer training using the VGG16 backbone, and intermediate-layer outputs are extracted as learned features. These features are then fed into a support vector regression (SVR) model to estimate beam position. On test data, the method achieves an 85.8% correct prediction rate, addressing noisy backgrounds and varying beam-spot appearances.","arXiv:2308.00718v1 [[physics.data-an](physics.data-an)] 1 Aug 2023  \nBeam Detection Based on Machine Learning Algorithms  \nHaoyuan Li, hyli16@stanofrd.edu Qing Yin, [qingyin@stanford.edu](qingyin@stanford.edu)  \nAugust 3, 2023  \nAbstract  \nThe positions of free electron laser beams on screens are precisely determined by a sequence of machine learning models. Transfer training is conducted in a self-constructed convolutional neural network based on VGG16 model. Output of intermediate layers are passed as features to a support vector regression model. With this sequence, 85.8% correct prediction is achieved on test data.  \nBeam Detection, SVM, CNN  \n1 Introduction  \nThe free electron laser(FEL) at Stanford Linear Accelerator Center(SLAC) is an ultra-fast X-ray laser. As one of the most advanced X-ray light source [5] [6], it is famous for its high brightness and short pulse duration: it is 10 billion times brighter than the world’s second brightest light source; the pulse duration is several tens femtoseconds.It plays a pivotal role in both fundamental science research and applied research [6] . The mechanism behind this laser is very delicate [5] . Thus to keep the laser in optimal working condition is challenging.The positions of the electron beams and the laser beams are of fundamental importance in the control and maintenance of this FEL.  \nCurrently, the task of locating beam spots heavily depends on human labor. This is mainly attributed to the wide varieties of beam spots and the presentation of strong noises as demonstrated in Figure 1, where the white square marks the boundary of the beam spot. Each picture requires a long sequence of signal processing methods to mark the beam position.  \nTo make things even worse, different instrument configurations and working conditions require different processing parameters. Within the frame of same configuration, the parameters will also drift away along with time advancement because of the inherent delicacy of the instrument. This makes tuning and maintaining the FEL a tedious and burdensome task for researchers.Currently, the data update frequency of the laser is 120 Hz. We can barely handle this. In 2020, after the scheduled update, the frequency will climb to 5000 Hz. Thus the only hope lies in automatic processing methods. Considering that simple signal processing method can not handle such nasty condition, We hope to come up with a general machine learning algorithm which is capable of locating the beams’ positions quickly and automatically. In this report, we demonstrate the consecutive application of neural network and supportive vector machine (SVM) to determine the positions of beam spots on virtual cathode camera (VCC) screen in simple cases.  \n2 Challenges and Strategies  \nA bird view of such a simple-looking task reveals the challenges behind it.  \n• Complicated background noises. The background noises are not static. In contrast, it is coherent in both time and space domain due to quantum effect. Thus we have to dynamically change the background when doing background subtraction for different images.  \n• Large variety in the intensity and shape of the beam spot. As is indicated above, the maximum intensity of the beam is incredibly high. However, it can also be 0 which corresponds to the case that no beam spot presented. The shape of the beam can also varies significantly from case to case.  \n• Ground truths is hard to find. The signal processing methods fail in finding the beam spot sometimes. There are also cases when the photo is so terrible that even human can not find the beam spot after all the signal processing procedures.  \nThe task is similar to face recognition and the expectation of solving it in one strike is unrealistic. Thus, we develop the following strategy. First, we develop an algorithm that is capable of dealing with cases where we have implemented signal processing to realize the fundamental functions. Second,  \nFigure 1: Raw photos  \nextend the a","cbCaisxfpafwycPM","https://ap.wps.com/l/cbCaisxfpafwycPM","pdf",1339104,1,12,"English","en",105,"# Introduction\n# Challenges and Strategies\n# Related Work\n# Dataset\n# Method\n## Pipeline\n## Feature Extraction","[{\"question\":\"Why is beam spot localization difficult for the FEL at SLAC?\",\"answer\":\"The task suffers from complex, time- and space-varying background noises, large variation in beam-spot intensity and shape, and limited availability of reliable ground-truth labels.\"},{\"question\":\"What machine learning pipeline is used to detect beam positions?\",\"answer\":\"A CNN acts as a feature extractor, where intermediate outputs are used as features to train a support vector regression (SVR) model for position estimation.\"},{\"question\":\"How is transfer training implemented and what performance is achieved?\",\"answer\":\"Transfer training is performed in a self-constructed convolutional neural network based on the VGG16 model, and the resulting sequence reaches 85.8% correct prediction on test data.\"}]","Beam Detection Based on Machine Learning Algorithms | PDF",1785679950,30,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"beam-detection-based-on-machine-learning-algorithms","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/beam-detection-based-on-machine-learning-algorithms/117843/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05","2026-08-02",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why is beam spot localization difficult for the FEL at SLAC?","Question",{"text":76,"@type":77},"The task suffers from complex, time- and space-varying background noises, large variation in beam-spot intensity and shape, and limited availability of reliable ground-truth labels.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What machine learning pipeline is used to detect beam positions?",{"text":81,"@type":77},"A CNN acts as a feature extractor, where intermediate outputs are used as features to train a support vector regression (SVR) model for position estimation.",{"name":83,"@type":74,"acceptedAnswer":84},"How is transfer training implemented and what performance is achieved?",{"text":85,"@type":77},"Transfer training is performed in a self-constructed convolutional neural network based on the VGG16 model, and the resulting sequence reaches 85.8% correct prediction on test data.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":46,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":46,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":29,"slug":122},"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]