[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119104-en":3,"doc-seo-119104-105":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},119104,5909877438554,"Maeve","https://ap-avatar.wpscdn.com/avatar/5600025385ad2bf12a7?_k=1778553567797529272",8,"Research & Report","The Management of Context in the Machine Learning Lifecycle","Novel techniques and systems are presented for managing data context throughout the machine learning (ML) lifecycle. Building on a 2018 vision, Flor and its evolutions—FlorDB and FlorDB with Build extensions—support comprehensive metadata capture and version control. An interview study is used to clarify how engineers operationalize the ML lifecycle, with emphasis on MLOps and iterative model development. Implementations in real-world applications for lawyers and journalists validate the value of rich data context and show how Application, Build, and Change contexts close the loop in ML.","UC Berkeley  \nUC Berkeley Electronic Theses and Dissertations  \nTitle  \nThe Management of Context in the Machine Learning Lifecycle  \nPermalink  \n[https://escholarship.org/uc/item/16g960sx](https://escholarship.org/uc/item/16g960sx)  \nAuthor  \nGarcia, Rolando  \nPublication Date  \n2024  \nPeer reviewed|Thesis/dissertation  \n[eScholarship.org](eScholarship.org) Powered by the California Digital Library  \nUniversity of California  \nThe Management of Context in the Machine Learning Lifecycle  \nBy  \nRolando Sanchez Garcia  \nA dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy  \nin  \nComputer Science  \nin the  \nGraduate Division  \nof the  \nUniversity of California, Berkeley  \nCommittee in charge:  \nProfessor Joseph M. Hellerstein, Chair Professor Joseph Gonzalez  \nProfessor Koushik Sen  \nAssociate Professor Fernando Perez  \nSummer 2024  \nThe Management of Context in the Machine Learning Lifecycle  \nCopyright 2024  \nby  \nRolando Sanchez Garcia  \n1  \nAbstract  \nThe Management of Context in the Machine Learning Lifecycle  \nby  \nRolando Sanchez Garcia  \nDoctor of Philosophy in Computer Science  \nUniversity of California, Berkeley  \nProfessor Joseph M. Hellerstein, Chair  \nWe present novel techniques and systems for managing data context within the machine learning (ML) lifecycle. Drawing from a vision laid out in 2018, we present Flor and its evolutions, FlorDB and FlorDB with Build extensions, designed for comprehensive metadata capture and version control in the ML lifecycle. A cornerstone of our approach is the use of an interview study to understand what the ML lifecycle is, and how engineers operationalize machine learning, focusing on MLOps and the iterative model development process. Through the implementation of these systems and their use in real-world applications for lawyersand journalists, we demonstrate the tangible benefits of rich data context in agile model development. In sum, we show how the integration of Application, Build, and Change contexts—The ABCs of Context—enables MLEs to close the loop in the ML lifecycle.  \ni  \nTo my grandfather, who worked as a citrus and cherry farmworker in California’s Central Valley despite being an academic at heart, and to countless talented others who toil and are toiling without prospects, so that some distant day their children and grandchildren may enjoy better opportunities. To my mother and father, who invested all that they could into my education, and taught me to prioritize school over work. To my wife, Victoria, and son, Joshua, who remind me every day what matters most, and show me the way forward.  \nii  \nContents  \nContents ii  \nList of Figures iv  \nList of Tables viii  \n1 Introduction 1  \n1.1 Build, Train, Deploy: What is the Machine Learning Lifecycle? ........ 1  \n1.2 Context: The Missing Piece in the Machine Learning Lifecycle ........ 3  \n1.3 Flor: The Management of Context in the Machine Learning Lifecycle .... 4  \n1.4 Flow with FlorDB ................................ 6  \n2 Operationalizing Machine Learning: An Interview Study 7  \n2.1 Introduction .................................... 7  \n2.2 Related Work ................................... 10  \n2.3 Methods ...................................... 15  \n2.4 Summary of Findings ............................... 21  \n2.5 Discussion ..................................... 36  \n2.6 Conclusion ..................................... 41  \n3 Context: The Missing Piece in the Machine Learning Lifecycle 42  \n3.1 Introduction .................................... 42  \n3.2 The Absence of Context ............................. 43  \n3.3 System Requirements for Model Lifecycle Management ............ 45  \n3.4 Related Work ................................... 48  \n3.5 Conclusion ..................................... 48  \n4 Application Context: Hindsight Logging for Model Training 50  \n4.1 Introduction .................................... 51  \n4.2 Methodical Hindsight Logging ................","cbCaiqZc60GviECY","https://ap.wps.com/l/cbCaiqZc60GviECY","pdf",37131107,1,154,"English","en",105,"# Introduction\n## Build, Train, Deploy: What is the Machine Learning Lifecycle?\n## Context: The Missing Piece in the Machine Learning Lifecycle\n## Flor: The Management of Context in the Machine Learning Lifecycle\n## Flow with FlorDB\n# Operationalizing Machine Learning: An Interview Study\n## Introduction\n## Related Work\n## Methods\n## Summary of Findings\n## Discussion\n## Conclusion\n# Context: The Missing Piece in the Machine Learning Lifecycle\n## Introduction\n## The Absence of Context\n## System Requirements for Model Lifecycle Management\n## Related Work\n## Conclusion\n# Application Context: Hindsight Logging for Model Training\n## Introduction\n## Methodical Hindsight Logging\n## Tooling for Hindsight Logging\n## Evaluation\n## Related Work\n## Conclusion\n# Change Over Time: Multiversion Hindsight Logging for Continuous Training\n## Introduction\n## Scenario: Catastrophic Forgetting\n## User Experience & API\n## System Architecture\n## Acquisitional Query Processing\n## Evaluation\n## Related Work\n## Conclusion\n# Context is All You Need: Closing the Loop in the ML Lifecycle\n## Introduction\n## Flow with FlorDB\n## Closing the Loop in the ML Lifecycle\n## Case Study\n## Discussion\n## Related Work\n## Conclusion\n# Conclusion\n## Looking Back\n## Next Steps","[{\"question\":\"What problem does the dissertation address in the ML lifecycle?\",\"answer\":\"It addresses how data context is missing or unmanaged across the ML lifecycle, affecting how engineering teams operationalize model development and MLOps practices.\"},{\"question\":\"What are Flor, FlorDB, and FlorDB with Build extensions?\",\"answer\":\"They are systems designed to capture comprehensive metadata and provide version control across ML lifecycle activities, supporting evolution of context management over time.\"},{\"question\":\"How does the work validate the benefits of rich data context?\",\"answer\":\"It demonstrates tangible benefits through real-world applications for lawyers and journalists, showing improved loop closure in agile and iterative model development.\"}]","The Management of Context in the Machine Learning Lifecycle | PDF",1785722407,388,{"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},"the-management-of-context-in-the-machine-learning-lifecycle","",{"@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/the-management-of-context-in-the-machine-learning-lifecycle/119104/",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 dissertation address in the ML lifecycle?","Question",{"text":75,"@type":76},"It addresses how data context is missing or unmanaged across the ML lifecycle, affecting how engineering teams operationalize model development and MLOps practices.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What are Flor, FlorDB, and FlorDB with Build extensions?",{"text":80,"@type":76},"They are systems designed to capture comprehensive metadata and provide version control across ML lifecycle activities, supporting evolution of context management over time.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the work validate the benefits of rich data context?",{"text":84,"@type":76},"It demonstrates tangible benefits through real-world applications for lawyers and journalists, showing improved loop closure in agile and iterative model development.","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"]