[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122918-en":3,"doc-seo-122918-105":30,"detail-sidebar-cat-0-en-105":95},{"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},122918,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Metadata Representations for Queryable Repositories of Machine Learning Models - paper overview","Machine learning organizations are building model repositories of pretrained models, commonly called model zoos, and rely on metadata to capture model and dataset properties. This metadata is essential for reporting, auditing, reproducibility, and interpretability, yet existing model zoos provide limited information and commonly used formats such as datasheets and model cards lack sufficient expressiveness. The paper proposes a unified metadata representation for model zoos, demonstrating that rich metadata supports search, reuse, comparison, and composition of machine learning models. It also presents the design and implementation of an advanced model zoo system based on the proposed representation.","Received 1 October 2023, accepted 27 October 2023, date of publication 6 November 2023, date of current version 13 November 2023. Digital Object Identifier 10.1109/ACCESS.2023.3330647  \nMetadata Representations for Queryable Repositories of Machine Learning Models  \nZIYU LI1,(Member, IEEE), HENK KANT 1, RIHAN HAI1,(Member, IEEE), ASTERIOS KATSIFODIMOS 1, MARCO BRAMBILLA2,(Member, IEEE), AND ALESSANDRO BOZZON1  \n1Department of Software Technology (ST), Faculty of Electrical Engineering Mathematics, and Computer Science (EEMCS), Delft University of Technology, 2628 CD Delft, The Netherlands  \n2Dipartimento di Elettronica Informazione e Bioingegneria (DEIB), Politecnico di Milano, 20133 Milan, Italy Corresponding author: Ziyu Li ([z.li-14@tudelft.nl](z.li-14@tudelft.nl))  \nThis work was supported in part by Cognizant.  \nABSTRACT Machine learning (ML) practitioners and organizations are building model repositories of pretrained models, referred to as model zoos. These model zoos contain metadata describing the properties of the ML models and datasets. The metadata serves crucial roles for reporting, auditing, ensuring reproducibility, and enhancing interpretability. Despite the growing adoption of descriptive formats like datasheets and model cards, the metadata available in existing model zoos remains notably limited. Moreover, existing formats have limited expressiveness, thus constraining the potential use of model repositories, extending their purpose beyond mere storage for pre-trained models. This paper proposes a unified metadata representation format for model zoos. We illustrate that comprehensive metadata enables a diverse range of applications, encompassing model search, reuse, comparison, and composition of ML models. We also detail the design and highlight the implementation of an advanced model zoo system built on top of our proposed metadata representation.  \nINDEX TERMS Machine learning, metadata representations, model zoo, model search.  \nI. INTRODUCTION  \nMachine learning (ML) is increasingly used across application domains such as video analytics [1], [2], autonomous driving [3], content moderation [4], traffic monitoring [5] and crowd detection [6] . While ML models can be (and often are) trained for specific purposes, there is a growing interest in reusing and re-purposing of pre-trained ML models [7] . This shift, motivated mainly by computational, economic, and environmental reasons, is evident from the proliferation of public, pre-trained ML model zoos, such as HuggingFace, Tensorflow Hub, and PyTorch Hub.1 These model zoos contain thousands of pre-trained models for diverse ML inference needs (e.g., recognition of classes/objects/concepts) . Thanks to model zoos, complex predictive and analytics tasks can benefit from reusing existing ML models.  \nThe associate editor coordinating the review of this manuscript and  \napproving it for publication was Adnan Kavak  .  \n1https://huggingface.co/, https://www.tensorflow.org/, https://pytorch. org/hub/  \nThe potential of model zoos is currently hindered by the lack of structured, comprehensive, and queryable metadata representations. Current repositories include a wide range of information, e.g., using model cards [8] . However, such information is mostly for human consumption, and the level of detail remains coarse-grained, thus preventing advanced repository automation and management functionalities. TABLE 1 presents the information provided by different public model zoos. The categories of the information cover different aspects of ML artifacts (e.g., model, dataset, performance) . We observe that current model zoos only provide limited information; for instance, PyTorch Hub provides only the ReadMe files from the source (e.g., a GitHub repository) . Insufficient information forces practitioners to search for additional metadata in external repositories and descriptive documents or repeatedly go through the ML lifecycle. These processes impede the reuse of mode","cbCaiaJJ31C9kM40","https://ap.wps.com/l/cbCaiaJJ31C9kM40","pdf",2831990,1,15,"English","en",105,"# Introduction\n## Motivation and limitations of existing model zoos\n## Need for structured, queryable metadata\n## Proposed unified metadata representation and applications\n# Related tools for ML lifecycle management\n## Standardization and integration challenges","[{\"question\":\"What problem does the paper address about existing model zoos?\",\"answer\":\"Existing model zoos lack structured, comprehensive, and queryable metadata, and available formats are often too limited for automation and advanced repository management.\"},{\"question\":\"What roles does metadata play in model repositories according to the paper?\",\"answer\":\"Metadata supports reporting, auditing, reproducibility, and interpretability, and it enables more advanced repository functions beyond simple storage.\"},{\"question\":\"What does the paper propose to improve model zoo usability?\",\"answer\":\"It proposes a unified metadata representation format for model zoos, designed to express rich metadata for ML models and datasets.\"},{\"question\":\"How can comprehensive metadata increase the value of model zoos?\",\"answer\":\"Comprehensive metadata enables model search, reuse, comparison, and composition, and it can power an advanced model zoo system built on the proposed representation.\"}]","Metadata Representations for Queryable Repositories of Machine Learning Models - paper overview | PDF",1785813663,38,{"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":90,"head_meta":92,"extra_data":94,"updated_unix":28},"metadata-representations-for-queryable-repositories-of-machine-learning-models-paper-overview","",{"@graph":36,"@context":89},[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/metadata-representations-for-queryable-repositories-of-machine-learning-models-paper-overview/122918/",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","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address about existing model zoos?","Question",{"text":75,"@type":76},"Existing model zoos lack structured, comprehensive, and queryable metadata, and available formats are often too limited for automation and advanced repository management.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What roles does metadata play in model repositories according to the paper?",{"text":80,"@type":76},"Metadata supports reporting, auditing, reproducibility, and interpretability, and it enables more advanced repository functions beyond simple storage.",{"name":82,"@type":73,"acceptedAnswer":83},"What does the paper propose to improve model zoo usability?",{"text":84,"@type":76},"It proposes a unified metadata representation format for model zoos, designed to express rich metadata for ML models and datasets.",{"name":86,"@type":73,"acceptedAnswer":87},"How can comprehensive metadata increase the value of model zoos?",{"text":88,"@type":76},"Comprehensive metadata enables model search, reuse, comparison, and composition, and it can power an advanced model zoo system built on the proposed representation.","https://schema.org",{"og:url":52,"og:type":91,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":93,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":46,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]