[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120513-en":3,"doc-seo-120513-105":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},120513,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Reinforcement learning-based architecture search for quantum machine learning - Abstract","Quantum machine learning models rely on encoding circuits to map input data into a quantum Hilbert space, yet these circuits are commonly selected heuristically even though their architecture strongly determines model behavior. This work introduces reinforcement learning to automatically generate problem-specific encoding circuits. A model-based RL strategy improves sample efficiency by reducing the number of circuit evaluations during search. The method employs a layered circuit structure to shrink the search space and handles multiple objectives including solution quality, hardware constraints, and circuit depth. Benchmarks across several datasets show clear performance gains versus reference, problem-agnostic, and classical baselines.","arXiv :2406 .02717v3 [ quant-ph] 5 Aug 2024  \nReinforcement learning-based architecture search for quantum  \nmachine learning  \nFrederic Rapp 1,2 , David A. Kreplin 1 , Marco F. Huber 1,2 , Marco Roth 1*  \n1 Fraunhofer Institute for Manufacturing Engineering and Automation IPA, Nobelstrasse 12, Stuttgart, 70569, Germany.  \n2 Institute of Industrial Manufacturing and Management IFF, University of Stuttgart, Allmandring 35, Stuttgart, 70569, Germany.  \n*Corresponding author(s). E-mail(s): [marco.roth@ipa.fraunhofer.de](marco.roth@ipa.fraunhofer.de) ; Contributing authors: frederic.rapp@ipa.fraunhofer.de;  \nAbstract  \nQuantum machine learning models use encoding circuits to map data into a quantum Hilbert space. While it is well known that the architecture of these circuits significantly influences core properties of the resulting model, they are often chosen heuristically. In this work, we present a novel approach using reinforcement learning techniques to generate problem-specific encoding circuits to improve the performance of quantum machine learning models. By specifically using a model-based reinforcement learning algorithm, we reduce the number of necessary circuit evaluations during the search, providing a sample-efficient framework. In contrast to previous search algorithms, our method uses a layered circuit structure that significantly reduces the search space. Additionally, our approach can account for multiple objectives such as solution quality, hardware restrictions and circuit depth. We benchmark our tailored circuits against various reference models, including models with problem-agnostic circuitsand classical models. Our results highlight the effectiveness of problem-specific encoding circuits in enhancing QML model performance.  \nKeywords: quantum computing, quantum machine learning, reinforcement learning, architecture search  \n1 Introduction  \nQuantum machine learning (QML) is a promising application of noisy intermediate-scale quantum computing (NISQ) [1], and has gained significant attention in recent years [2] . Central to many QML approaches is the use of encoding circuits, which map data into a quantum Hilbert space. These circuits may optionally be parameterized and are crucial components of methods such as quantum neural networks (QNN) [3, 4] and quantum kernel methods [5–7] .  \nUnlike variational algorithms used for quantum simulation [8, 9] or optimization [10], there are few restrictions when it comes to the design and structure of quantum circuits used for QML. They can be tailored to accommodate hardware-specific constraints such as connectivity or native gates.  \nThis design freedom allows for the creation of circuits that are well-suited to specific hardware and problem constraints. However, if the encoding circuit is not chosen properly, optimization can become infeasible [11, 12] and poorly chosen circuits can result in inadequate models [13] . While  \nthere are some established principles such as using data re-uploading and hardware efficient gate sets, they may not necessarily maximize the model’s performance. From the no free lunch theorem [14] it is well known that optimal performance is achieved by a problem specific model. Although this has been confirmed for QML models [13, 15], there are no established guidelines for how to construct QML encoding circuits that optimize the performance of the resulting model for a given task.  \nAs a remedy, recent work has focused on incorporating properties of the data, such as symmetries, into QML architectures to provide models with a informed inductive bias [16–19] . Nonetheless, many of these approaches require expert knowledge of problem domains and quantum circuit design, a prerequisite shared by other informed design choices [20] . This gap motivates the development of automated methods that can create and evaluate different encoding circuits, capable of generating problem-specific encoding circuits from input data.  \nIn this work, we introduce a","cbCaifNKA60b04SS","https://ap.wps.com/l/cbCaifNKA60b04SS","pdf",903388,1,14,"English","en",105,"# Introduction\n## Motivation and background\n## Reinforcement learning approach (MuZero)\n## Benchmarking and contributions\n## Document structure","[{\"question\":\"Why does the encoding-circuit architecture matter in quantum machine learning?\",\"answer\":\"Encoding circuits map data into a quantum Hilbert space, and their architecture strongly influences key properties and ultimately the model performance.\"},{\"question\":\"How does the proposed method generate encoding circuits?\",\"answer\":\"It uses model-based reinforcement learning (focused on MuZero) where observations represent the current circuit, actions add the next gate layer, and rewards come from cross-validation training performance.\"},{\"question\":\"What advantages does the layered circuit structure provide?\",\"answer\":\"It reduces the effective search space, making the architecture search more efficient while enabling constraints such as solution quality, hardware restrictions, and circuit depth to be considered.\"}]","Reinforcement learning-based architecture search for quantum machine learning - Abstract | PDF",1785730431,35,{"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},"reinforcement-learning-based-architecture-search-for-quantum-machine-learning-abstract","",{"@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/reinforcement-learning-based-architecture-search-for-quantum-machine-learning-abstract/120513/",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},"Why does the encoding-circuit architecture matter in quantum machine learning?","Question",{"text":75,"@type":76},"Encoding circuits map data into a quantum Hilbert space, and their architecture strongly influences key properties and ultimately the model performance.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the proposed method generate encoding circuits?",{"text":80,"@type":76},"It uses model-based reinforcement learning (focused on MuZero) where observations represent the current circuit, actions add the next gate layer, and rewards come from cross-validation training performance.",{"name":82,"@type":73,"acceptedAnswer":83},"What advantages does the layered circuit structure provide?",{"text":84,"@type":76},"It reduces the effective search space, making the architecture search more efficient while enabling constraints such as solution quality, hardware restrictions, and circuit depth to be considered.","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"]