[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125946-en":3,"doc-seo-125946-105":31,"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":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},125946,137451207643,"Noah","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","Hamiltonian Learning using Machine Learning Models Trained with Continuous Measurements - arXiv","Hamiltonian Learning using Machine Learning models trained with continuous weak measurements focuses on estimating Hamiltonian parameters of qubits from measurement records without requiring explicit quantum-state representation. The work studies supervised training using labeled records and an unsupervised approach without labels, where a physical integrator is combined with a recurrent neural network for model-free correction. Tests on two-qubit systems show accurate parameter prediction, improved performance with larger training sets, and robustness under unanticipated single-particle relaxation errors.","arXiv :2404 .05526v1 [ quant-ph] 8 Apr 2024  \nSAND2024-04128O  \nHamiltonian Learning using Machine Learning Models Trained with Continuous  \nMeasurements  \nKris Tucker, 1, ∗ Amit Kiran Rege,2,† Conor Smith,3, 4 Claire Monteleoni,2, 5 and Tameem Albash6  \n1 Department of Applied Math, University of Colorado Boulder, USA  \n2 Department of Computer Science, University of Colorado Boulder, USA  \n3 Center for Quantum Information and Control, University of New Mexico, University of New Mexico, Albuquerque, NM, USA  \n4 Department of Electrical and Computer Engineering, University of New Mexico, Albuquerque, NM, USA  \n5 INRIA Centre de Recherche de Paris, France  \n6 Center for Computing Research, Sandia National Laboratories, Albuquerque, New Mexico 87185, USA  \nWe build upon recent work on using Machine Learning models to estimate Hamiltonian parameters using continuous weak measurement of qubits as input. We consider two settings for the training  \nof our model: (1) supervised learning where the weak measurement training record can be labeled  \nwith known Hamiltonian parameters, and (2) unsupervised learning where no labels are available.  \nThe first has the advantage of not requiring an explicit representation of the quantum state, thus  \npotentially scaling very favorably to larger number of qubits. The second requires the implementa  \ntion of a physical model to map the Hamiltonian parameters to a measurement record, which we  \nimplement using an integrator of the physical model with a recurrent neural network to provide a  \nmodel-free correction at every time step to account for small effects not captured by the physical  \nmodel. We test our construction on a system of two qubits and demonstrate accurate prediction  \nof multiple physical parameters in both the supervised and unsupervised context. We demonstrate  \nthat the model benefits from larger training sets establishing that it is in fact “learning,” and we  \nshow robustness to errors in the assumed physical model by achieving accurate parameter estimation  \nin the presence of unanticipated single particle relaxation.  \nI. INTRODUCTION  \nAs the scale, complexity, and availability of quantum devices continues to grow over the coming decades [1], the ability to accurately characterize device parameters, especially unwanted effects, will become ever-more important. For example, non-local errors such as crosstalkin large systems [2, 3] can be problematic for quantum error correction methods [4–8] . Detection of these unwanted effects is a critical first step in mitigating them, with efforts to do so complicated by the exponential scaling of the dimension of the Hilbert space with system size and the challenges of modeling increasingly complex systems.  \nRecent advances in the application of machine learning (ML) tools to quantum systems have shown promise in overcoming some of these difficulties. Scalable ML models have been used to represent quantum states for state tomography [9–14] and evolution [14, 15], for learning unknown dynamics [16–18], as well as for learning device characterization from measurements [19–23] . Increasingly, a priori knowledge of a physical system has been combined with ML models to improve accuracy and interpretability [24] . The model-free nature of many ML solutions has made them scalable and robust to many of the common pitfalls of physical models, such as nonMarkovian dynamics [15, 19–21] . There are trade-offs,  \n∗ [kristopher.tucker@colorado.edu](kristopher.tucker@colorado.edu)[ ](kristopher.tucker@colorado.edu)† [Amit.Rege@colorado.edu](Amit.Rege@colorado.edu)  \nhowever, as more abstract models are generally less interpretable, which limits the physical insights that can begleaned from them. Finding a balance between these two competing properties, representability and interpretability, thus becomes an important challenge.  \nIn this work, we build on recent advances in device characterization applying ML models to the continuous measurement of","cbCaicBdJpPAxDlS","https://ap.wps.com/l/cbCaicBdJpPAxDlS","pdf",739774,4,1,12,"English","en",105,"# Introduction\n## Quantum device characterization and challenges\n## Machine learning for quantum systems\n## This work: supervised and unsupervised Hamiltonian learning","[{\"question\":\"What inputs and targets does the proposed model use for Hamiltonian parameter learning?\",\"answer\":\"The model uses continuous weak measurement records as inputs and targets Hamiltonian parameters. In supervised learning, measurement records are paired with known parameters; in unsupervised learning, labels are not required.\"},{\"question\":\"How does the unsupervised training approach work without labeled Hamiltonian parameters?\",\"answer\":\"It compares the input measurement record with one reconstructed from parameter estimates. A layer combines a physical-model integrator with a recurrent neural network to apply model-free corrections at each time step.\"},{\"question\":\"What advantages are claimed for supervised learning in this framework?\",\"answer\":\"Supervised learning is independent of any explicit quantum-state representation and can scale better with more qubits. After training, parameter estimation can be performed quickly, including for parameters not encountered during training.\"}]","Hamiltonian Learning using Machine Learning Models Trained with Continuous Measurements - arXiv | PDF",1785902174,30,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"hamiltonian-learning-using-machine-learning-models-trained-with-continuous-measurements-arxiv","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,52],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":51},"https://docshare.wps.com/document/research-report/",3,{"item":53,"name":13,"@type":44,"position":20},"https://docshare.wps.com/document/hamiltonian-learning-using-machine-learning-models-trained-with-continuous-measurements-arxiv/125946/",{"url":53,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-22","2026-08-05",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},"What inputs and targets does the proposed model use for Hamiltonian parameter learning?","Question",{"text":76,"@type":77},"The model uses continuous weak measurement records as inputs and targets Hamiltonian parameters. In supervised learning, measurement records are paired with known parameters; in unsupervised learning, labels are not required.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the unsupervised training approach work without labeled Hamiltonian parameters?",{"text":81,"@type":77},"It compares the input measurement record with one reconstructed from parameter estimates. A layer combines a physical-model integrator with a recurrent neural network to apply model-free corrections at each time step.",{"name":83,"@type":74,"acceptedAnswer":84},"What advantages are claimed for supervised learning in this framework?",{"text":85,"@type":77},"Supervised learning is independent of any explicit quantum-state representation and can scale better with more qubits. After training, parameter estimation can be performed quickly, including for parameters not encountered during training.","https://schema.org",{"og:url":53,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":53},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":30,"slug":122},"research-report",{"id":124,"doc_module":4,"doc_module_name":47,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":47,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":47,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":47,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]