[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119779-en":3,"doc-seo-119779-105":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},119779,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","Coreset selection can accelerate quantum machine learning models with provable generalization","Quantum neural networks and quantum kernels face a major bottleneck in training efficiency when working with large datasets, where QNN training lacks efficient primitives like back-propagation and quantum-kernel training requires constructing an O(N^2) kernel matrix via repeated quantum circuit evaluations. The work introduces a unified coreset selection method that distills a carefully chosen subset from the full training set, then analyzes generalization error bounds for QNNs and quantum kernels trained on these coresets, showing comparable performance to full-data training while reducing training cost. Numerical simulations demonstrate acceleration across tasks such as synthetic classification, quantum correlation identification, and quantum compiling.","arXiv :2309 . 10441v1 [ quant-ph] 19 Sep 2023  \nCoreset selection can accelerate quantum machine learning models with provable generalization  \nYiming Huang 1,2 , Huiyuan Wang3 , Yuxuan Du4 , and Xiao Yuan 1,2  \n1 Center on Frontiers of Computing Studies, Peking University, Beijing 100871, China 2 School of Computer Science, Peking University, Beijing 100871, China 3 Peterhouse, Univeristy of Cambridge, CB2 1RD Cambridge, Cambridgeshire, U. K.  \n4 JD Explore Academy, Beijing 101111, China  \nQuantum neural networks (QNNs) and quantum kernels stand as prominent figures in the realm of quantum machine learning, poised to leverage thenascent capabilities of near-term quantum computers to surmount classical machine learning challenges. Nonetheless, the training efficiency challenge poses a limitation on both QNNs and quantum kernels, curbing their efficacy when applied to extensive datasets. To confront this concern, we present a unified approach: coreset selection, aimed at expediting the training of QNNsand quantum kernels by distilling a judicious subset from the original training dataset. Furthermore, we analyze the generalization error bounds of QNNsand quantum kernels when trained on such coresets, unveiling the comparable performance with those training on the complete original dataset. Through systematic numerical simulations, we illuminate the potential of coreset selection in expediting tasks encompassing synthetic data classification, identification of quantum correlations, and quantum compiling. Our work offers a useful way to improve diverse quantum machine learning models with a theoretical guarantee while reducing the training cost.  \n1 Introduction  \nQuantum neural networks (QNNs) [1–4] and quantum kernels [5 , 6] have emerged as pivotal models in the burgeoning field of quantum machine learning (QML) [7–9], poised to unlock the power of near-term quantum computers to address challenges that elude classical machine learning paradigms [10 , 11] . The allure of these models is rooted in a fusion of theoretical advances and practical adaptability. That is, theoretical evidence showcases their superiority over classical counterparts in diverse scenarios, spanning synthetic datasets, discrete logarithmic problems, and quantum information processing tasks [6 , 12–16], as measured by sample complexity and runtime considerations. Complementing their theoretical strength, their implementation displays flexibility, adeptly accommodating constraints posed by contemporary quantum hardware, including qubit connectivity and limited circuit depth. This convergence of theoretical promise and practical flexibility has spurred a wave of empirical investigations, substantiating the viability and potential benefits of QNNs and quantum kernels across real-world applications such as computer vision [17–19] and quantum physics [20 , 21] .  \nDespite their promising potential, QNNs and quantum kernels confront a pertinent challenge concerning the training efficiency, resulting in a constrained practical applicability towards large-scale datasets [22] . This limitation is particularly evident due to the absence of fundamental training mechanisms like back-propagation and batch gradient descent in the majority of QNNs, imperative for the swift training of deep neural networks [23] . Similarly, the training process of quantum kernels necessitates the collection of a kernel matrix of size O (N2 ) , with N being the number of training examples and each entry demanding independent evaluation via a specific quantum circuit. Consequently, the capacities of both QNNs and quantum kernels to effectively navigate vast training datasets, characterized by millions of data points, are compromised.  \nIn response to the above challenge, several research lines have emerged to enhance the training efficiency of QNNs. The first line embarks on improving the optimizer or the initialization methods, seeking to expedite convergence towards the minimal empirical r","cbCaisfLFG9qbUYt","https://ap.wps.com/l/cbCaisfLFG9qbUYt","pdf",1323821,1,25,"English","en",105,"# Introduction\n## Training efficiency challenges for QNNs and quantum kernels\n## Prior approaches to improve training efficiency\n## Introducing coreset selection and theoretical guarantees","[{\"question\":\"What problem does the paper address in quantum machine learning training?\",\"answer\":\"It addresses the limited training efficiency of quantum neural networks and quantum kernels on large datasets, including lack of standard deep-learning training mechanisms for QNNs and the O(N^2) cost of building quantum kernel matrices.\"},{\"question\":\"How does coreset selection improve training efficiency?\",\"answer\":\"It speeds up training by distilling a smaller, well-chosen subset (a coreset) from the original training dataset, then training the QNNs or quantum kernels on this reduced set.\"},{\"question\":\"What theoretical result is provided for models trained on coresets?\",\"answer\":\"The paper derives generalization error bounds for QNNs and quantum kernels trained on coresets, showing performance comparable to training on the full original dataset.\"}]","Coreset selection can accelerate quantum machine learning models with provable generalization | PDF",1785726271,63,{"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},"coreset-selection-can-accelerate-quantum-machine-learning-models-with-provable-generalization","",{"@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/coreset-selection-can-accelerate-quantum-machine-learning-models-with-provable-generalization/119779/",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","2026-08-03",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},"What problem does the paper address in quantum machine learning training?","Question",{"text":75,"@type":76},"It addresses the limited training efficiency of quantum neural networks and quantum kernels on large datasets, including lack of standard deep-learning training mechanisms for QNNs and the O(N^2) cost of building quantum kernel matrices.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does coreset selection improve training efficiency?",{"text":80,"@type":76},"It speeds up training by distilling a smaller, well-chosen subset (a coreset) from the original training dataset, then training the QNNs or quantum kernels on this reduced set.",{"name":82,"@type":73,"acceptedAnswer":83},"What theoretical result is provided for models trained on coresets?",{"text":84,"@type":76},"The paper derives generalization error bounds for QNNs and quantum kernels trained on coresets, showing performance comparable to training on the full original dataset.","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":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]