[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117791-en":3,"doc-seo-117791-105":30,"detail-sidebar-cat-0-en-105":91},{"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":4,"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},117791,687197100911,"Himbo","https://ap-avatar.wpscdn.com/avatar/a000239b6f1da00475?x-image-process=image/resize,m_fixed,w_180,h_180&k=1785132997149421697",8,"Research & Report","Applying Machine Learning Methods to Laser Acceleration of Protons - Lessons Learned from Synthetic Data","Neural networks and other machine learning algorithms are increasingly used in ultra-intense laser physics, where experiments are often data-limited by low shot rates. This project investigates whether machine learning remains accurate when trained on small synthetic datasets. Using synthetic data derived from a Fuchs et al. 2006 model, it evaluates three learning methods, including a two-hidden-layer neural network, for speed and accuracy. Results support training strategies for neural networks on real experimental data.","Applying Machine Learning Methods to Laser Acceleration of Protons: Lessons Learned from Synthetic Data  \nPresented in Partial Fulfillment of the Requirements for graduation ”with Honors Research Distinction in Physics” in the undergraduate colleges of The Ohio State  \nUniversity  \nby  \nTom Zhang  \nThe Ohio State University  \nApril 2023  \nProject Advisor: Professor Chris Orban, Department of Physics  \nAbstract  \nNeural networks and other machine learning algorithms are beginning to be used in ultra-intense laser physics. Compared to other machine learning applications, ultra-intense laser systems are typically data poor because of limitations on the number of shots per second. An important concern is that the machine learning method being used remains accurate even when trained on a relatively small number of data points. By using synthetic data based on a model proposed by Fuchs et al. 2006, we seek to explore the speed and accuracy of three different machine learning methods including a neural network with two hidden layers. The results of this project can potentially be applied to training neural networks on real experimental data sets.  \nContents  \n1 Introduction 1  \n2 Experimental Description 2  \n2.1 The Dataset ................................... 2  \n2.2 Network Architecture .............................. 2  \n2.3 Neural Network Optimization ......................... 4  \n2.4 Neural Network Comparison with Other Machine Learning Methods .... 4  \n3 Results 6  \n3.1 Neural Network Performance .......................... 6  \n3.2 Performance Compared to Other Machine Learning Algorithms ...... 9  \n4 Conclusion 13  \n5 Acknowledgements 14  \n6 Works Cited 15  \n1 Introduction  \nThe field of ultra-intense laser physics features many non-linear physical systems in which it is either difficult to analytically solve, or computationally expensive to simulate. As such, laser physicists are increasingly turning to machine learning techniques to optimize their laser systems. For example, a recent paper by Loughran et al. (2023) used Gaussian Process regression to conduct Bayesian optimization to optimize the maximum proton energy of their laser pulses automatically[1] .  \nThere are questions, however, over how easily machine learning models can be applied to laser systems in general. While Loughran et al. saw success in using machine learning to optimize their laser, the laser system that Loughran et al. used had operated at 1 Hz. Laser systems with kHz repetition rates may produce more data than some machine learning methods may be able to process in quasi-real time.  \nFurthermore, there are many different machine learning models that can be utilized. Bethke et al (2021), for example, had used invertible neural networks to predict the result of Laser-Wakefield Acceleration simulations[2] . Comparing the performance of different machine learning models with each other may better aid laser physicists into selecting the appropriate machine learning model given their needs. Thus, the goal of this thesis is two-fold: We will use a simple, fully connected neural network and explore its performance given different constraints to its training length and data set size, and we will also compare the performance of this neural network to other machine learning models.  \n2 Experimental Description  \n2.1 The Dataset  \nWe use synthetic data that was primarily based on a physical model introduced by Fuchs et al. (2005)[3] . Some modifications to the relationship between the laser period and acceleration timescale for protons were introduced to the model based on the work done by Djordjevic et al (2021)[4] and to match the model to the experimental results found by Morrison et al. (2018)[5] In this model, we have five input parameters: Laser intensity, pulse duration, target thickness, focal distance, and spot size, which combine to result in the maximum proton energy as well as the number of protons generated. This model can be further extended to calcula","cbCaisvDyUuUxwrO","https://ap.wps.com/l/cbCaisvDyUuUxwrO","pdf",744663,1,19,"English","en",105,"# Introduction\n## Experimental context and motivation\n# Experimental Description\n## The Dataset\n## Network Architecture\n## Neural Network Optimization\n## Neural Network Comparison with Other Machine Learning Methods\n# Results\n## Neural Network Performance\n## Performance Compared to Other Machine Learning Algorithms\n# Conclusion\n# Acknowledgements\n# Works Cited","[{\"question\":\"Why does the project focus on synthetic data for laser proton acceleration?\",\"answer\":\"Ultra-intense laser systems are often data-poor because of limited shots per second. Synthetic data provides a controlled way to test machine learning accuracy under small training set sizes.\"},{\"question\":\"What inputs does the synthetic dataset use to predict proton outcomes?\",\"answer\":\"The model uses five input parameters—laser intensity, pulse duration, target thickness, focal distance, and spot size—to determine maximum proton energy and the number of protons generated.\"},{\"question\":\"Which machine learning methods are compared in this work?\",\"answer\":\"The study evaluates three methods, including a fully connected neural network with two hidden layers, and compares its performance against other machine learning algorithms using relative error metrics.\"}]","Applying Machine Learning Methods to Laser Acceleration of Protons - Lessons Learned from Synthetic Data | PDF",1785679592,48,{"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":86,"head_meta":88,"extra_data":90,"updated_unix":28},"applying-machine-learning-methods-to-laser-acceleration-of-protons-lessons-learned-from-synthetic-data","",{"@graph":36,"@context":85},[37,54,68],{"@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/applying-machine-learning-methods-to-laser-acceleration-of-protons-lessons-learned-from-synthetic-data/117791/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why does the project focus on synthetic data for laser proton acceleration?","Question",{"text":75,"@type":76},"Ultra-intense laser systems are often data-poor because of limited shots per second. Synthetic data provides a controlled way to test machine learning accuracy under small training set sizes.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What inputs does the synthetic dataset use to predict proton outcomes?",{"text":80,"@type":76},"The model uses five input parameters—laser intensity, pulse duration, target thickness, focal distance, and spot size—to determine maximum proton energy and the number of protons generated.",{"name":82,"@type":73,"acceptedAnswer":83},"Which machine learning methods are compared in this work?",{"text":84,"@type":76},"The study evaluates three methods, including a fully connected neural network with two hidden layers, and compares its performance against other machine learning algorithms using relative error metrics.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"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":21,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]