[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117346-en":3,"doc-seo-117346-105":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},117346,1099514068035,"Ezra","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Machine learning with requirements - A manifesto - Pyramid development process","Machine learning has achieved major breakthroughs across domains, yet deploying it in high-stakes and safety-critical settings remains challenging because models can be brittle and unreliable. The work argues that requirements definition and satisfaction can make models better aligned with real-world needs, highlighting cases where requirements emerge naturally and where ignoring them leads to severe consequences. A pyramid development process integrates requirements specification throughout the machine-learning pipeline, enabling continual mutual influence between requirements and subsequent phases. The paper further emphasizes neuro-symbolic AI as a key enabler for this integration, supporting safe and effective deployment in sensitive areas.","Machine learning with requirements: A manifesto  \nEleonora Giunchiglia 1 , Fergus Imrie2 ,  \nMihaela van der Schaar3,4 and Thomas Lukasiewicz5,6  \nNeurosymbolic Artificial Intelligence  \nVolume 1: 1–12 © 2025 – The authors. Published by IOS  \nPress.  \nArticle reuse guidelines:  \n[sagepub.com/journals-permissions](sagepub.com/journals-permissions)[ ](sagepub.com/journals-permissions)DOI: 10.3233/NAI-240767 [journals.sagepub.com/home/nai](journals.sagepub.com/home/nai)  \nEditor: Annette Ten Teije, Vrije Universiteit Amsterdam, The Netherlands  \nSolicited reviews: Floris van der Hengst, Vrije Universiteit Amsterdam, The Netherlands; two anonymous reviewers  \nAbstract  \nIn the recent years, machine learning has made great advancements that have been at the root of many breakthroughsin different application domains. However, it is still an open issue how to make them applicable to high-stakes or safetycritical application domains, as they can often be brittle and unreliable. In this paper, we argue that requirements definition and satisfaction can go a long way to make machine learning models even more fitting to the real world, especially in critical domains. To this end, we present two problems in which (i) requirements arise naturally, (ii) machine learning models are or can be fruitfully deployed, and (iii) neglecting the requirements can have dramatic consequences. Our proposed pyramid development process integrates requirements specification into every stage of the machine learning pipeline, ensuring mutual influence between requirements and subsequent phases. Additionally, we explore the pivotal role of Neuro-symbolic AI in facilitating this integration, paving the way for more reliable and robust machine learning applications in critical domains. Through this approach, we aim to bridge the gap between theoretical advancements and practical implementations, ensuring machine learning’s safe and effective deployment in sensitive areas.  \nKeywords  \nSafe AI, machine learning, requirements, machine learning operations, software engineering  \nReceived: 10 July 2023; revised: 24 June 2024; accepted: 8 July 2024  \n1 Introduction  \nIn recent years, machine learning has made great advancements that have been at the root of many breakthroughs in different application domains. For example, AlphaFold [63] is a deep learning model that solved the “protein folding problem”, a grand challenge in the field of biology for more than half a century, while Halicin [71] is the first antibiotic discovered using machine learning, which could help in the battle against bacterial resistance. Results like those above, though ground-breaking, overshadow the dangers that come with the careless use of machine learning models in critical applications. Indeed, even though such models report an astonishingly high performance in terms of accuracy (or alternatively  \n1 Imperial-X, Department of Electrical and Electronic Engineering, Imperial College, United Kingdom  \n2 Department of Electrical and Computer Engineering, University of California, Los Angeles, USA  \n3 DAMTP, University of Cambridge, United Kingdom  \n4Alan Turing Institute, United Kingdom  \n5 Institute of Logic and Computation, Vienna University of Technology, Austria  \n6 Department of Computer Science, University of Oxford, United Kingdom  \nCorresponding author:  \nEleonora Giunchiglia, Imperial-X, Department of Electrical and Electronic Engineering, Imperial College, United Kingdom.  \nEmail: [e.giunchiglia@imperial.ac.uk](e.giunchiglia@imperial.ac.uk)  \n Creative Commons CC BY: This is an open access article distributed under the terms of the Creative Commons Attribution (CC BY 4.0) License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is  \nproperly cited.  \nchosen metric), they do not give any guarantee that the model will not have any unintended behaviour when used in practice. Indeed, as pointed out by Rudin in [59], there have been severa","cbCaip24Ta8dtb16","https://ap.wps.com/l/cbCaip24Ta8dtb16","pdf",602467,1,12,"English","en",105,"# Introduction\n## Machine learning breakthroughs and safety risks\n## Requirements specification and verification\n## Two motivating problem settings\n## Pyramid development process\n## Role of neuro-symbolic AI","[{\"question\":\"Why is requirements work important for machine learning in safety-critical domains?\",\"answer\":\"The document explains that unintended behaviors arise when models violate requirements that may be known before data collection and development. Integrating requirements reduces the risk of dangerous unintended outcomes.\"},{\"question\":\"What is the proposed pyramid development process?\",\"answer\":\"It integrates requirements specification into every stage of the machine learning pipeline. This creates mutual influence between requirements and later pipeline phases.\"},{\"question\":\"How does neuro-symbolic AI support the integration of requirements and ML development?\",\"answer\":\"The paper presents neuro-symbolic AI as pivotal for facilitating the requirements integration process, helping enable more reliable and robust machine learning for critical domains.\"}]","Machine learning with requirements - A manifesto - Pyramid development process | PDF",1785675295,30,{"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},"machine-learning-with-requirements-a-manifesto-pyramid-development-process","",{"@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/machine-learning-with-requirements-a-manifesto-pyramid-development-process/117346/",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","2026-08-02",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 is requirements work important for machine learning in safety-critical domains?","Question",{"text":75,"@type":76},"The document explains that unintended behaviors arise when models violate requirements that may be known before data collection and development. Integrating requirements reduces the risk of dangerous unintended outcomes.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the proposed pyramid development process?",{"text":80,"@type":76},"It integrates requirements specification into every stage of the machine learning pipeline. This creates mutual influence between requirements and later pipeline phases.",{"name":82,"@type":73,"acceptedAnswer":83},"How does neuro-symbolic AI support the integration of requirements and ML development?",{"text":84,"@type":76},"The paper presents neuro-symbolic AI as pivotal for facilitating the requirements integration process, helping enable more reliable and robust machine learning for critical domains.","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},[93,97,101,105,110,115,120,122,127,130,134],{"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":29,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]