[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124172-en":3,"doc-seo-124172-105":30,"detail-sidebar-cat-0-en-105":90},{"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},124172,1374391975076,"Riley","https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051",6,"Technology","ORES-Inspect - A technology probe for machine learning audits on enwiki","Auditing the machine learning models used on Wikipedia is essential to keep vandalism-detection processes fair and effective, yet audits are hard because stakeholders differ in priorities and assembling evidence for a model’s (in)efficacy is technically complex. An interface was designed to help editors learn about and audit the ORES edit quality model. ORES-Inspect is an open-source web tool and a technology probe studying how editors think about auditing the many ML models deployed on Wikipedia.","Wiki Workshop (11th edition)– June 20, 2024  \nORES-Inspect: A technology probe for machine learning audits on enwiki  \nZachary Levonian  \nDigital Harbor Foundation  \nLauren Hagen  \nUniversity of Minnesota  \nLu Li  \nUniversity of Pennsylvania  \nJada Lilleboe  \nUniversity of Minnesota  \nSolvejg Wastvedt  \nUniversity of Minnesota  \nAaron Halfaker  \nMicrosoft Research  \nLoren Terveen  \nUniversity of Minnesota  \narXiv :2406 .08453v1 [ cs .HC] 12 Jun 2024  \nAbstract  \nAuditing the machine learning (ML) models used on Wikipedia is important for ensuring that vandalism-detection processes remain fair and effective. However, conducting audits is challenging because stakeholders have diverse priorities and assembling evidence for a model’s [in]efficacy is technically complex. We designed an interface to enable editors to learn about and audit the performance of the ORES edit quality model. ORES-Inspect1 is an open-source web tool and a provocative technology probe for researching how editors think about auditing the many ML models used on Wikipedia. We describe the design of ORES-Inspect and our plans for further research with this system.  \nKeywords: machine learning, auditing, tools, ORES, edit quality  \nIntroduction  \nORES is a widely-used service for building and hosting machine learning models requested by the Wikipedia community (Halfaker and Geiger, 2020) . Of particular relevance is the edit quality model, which makes predictions about the quality of individual Wikipedia edits and is used in other systems for vandalism detection and removal. ORES’ edit quality predictions directly influence the likelihood of an edit being reverted (TeBlunthuis et al., 2020) . This impact is a notable success for communitycentered and participatory machine learning processes: ORES is hosting an increasing number of models.2  \nA key challenge for ORES and other machine learning services is that it is hard to determine if a model is consistently producing reasonable outputs. In other words, it is hard to audit these models. There are many barriers to auditing complex machine learning systems like ORES:(a) identifying a relevant sample of incorrect predictions,(b) determining if those incorrect predictions represent a pattern of undesired behavior (a “bug”), and (c) convincing system designers to fix the undesired behavior. To  \n1[https://ores-inspect.toolforge.org](https://ores-inspect.toolforge.org)  \n2ORES is being replaced with LiftWing, but this work is applicable to any revscoring model.  \naddress those barriers, we are building ORES-Inspect, an open-source3 tool to audit the behavior of the ORES edit quality model for English Wikipedia.  \nIn the consensus-driven Wikipedia context, the developers of ML-driven systems like ORES are enthusiastic about receiving community input on problems or potential areas for improvement. Thus, the key design objective for ORES-Inspect is to address problems (a) and (b) by making it easy to identify high-quality quantitative evidence of the ORES edit quality model’s behaviors. Weare developing ORES-Inspect as a “technology probe”to reflect on the process of conducting ML audits in the Wikipedia context by highlighting the benefits and challenges of collecting quantitative evidence of system bugs (Hutchinson et al., 2003) .  \nFunctionally, ORES-Inspect is a labeling interface for individual Wikipedia edits. The key intuition is that any Wikipedia user maybe interested in auditing a system like ORES, but different auditors will have different priorities (e.g. are new editors unfairly targeted, is vandalism on stubs missed more often than on larger articles, etc.) . For that reason, the process of auditing is the process of quantifying one’s intuitions and identifying evidence that a single misclassification represents a pattern that should be changed. Therefore, we designed ORES-Inspect as a provocation: it is designed to educate editors about how ML models can be audited and how to translate intuitions into hig","cbCaidjy74YvVjNh","https://ap.wps.com/l/cbCaidjy74YvVjNh","pdf",622604,1,4,"English","en",105,"# Introduction\n## ORES and the need for auditing\n## Challenges in conducting ML audits\n## Designing ORES-Inspect as a technology probe\n## Labeling interface and auditing intent\n## Interface structure: four activity phases","[{\"question\":\"Why is auditing the Wikipedia machine learning models important?\",\"answer\":\"It helps ensure vandalism-detection systems remain fair and effective. ORES’ edit quality predictions affect whether edits are likely to be reverted, so audit evidence can target unfair or incorrect behavior.\"},{\"question\":\"What makes auditing complex ML systems difficult in this context?\",\"answer\":\"Auditors must find relevant incorrect predictions, determine whether they reflect a consistent bug-like pattern, and persuade system designers to change the undesired behavior.\"},{\"question\":\"What is ORES-Inspect and what is its main purpose?\",\"answer\":\"ORES-Inspect is an open-source web tool that enables editors to audit the ORES edit quality model. It functions as a technology probe to study how editors translate intuitions into quantitative evidence about model behavior.\"}]","ORES-Inspect - A technology probe for machine learning audits on enwiki | PDF",1785820849,10,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":28},"ores-inspect-a-technology-probe-for-machine-learning-audits-on-enwiki","",{"@graph":36,"@context":84},[37,53,67],{"@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/technology/",3,{"item":52,"name":13,"@type":43,"position":21},"https://docshare.wps.com/document/ores-inspect-a-technology-probe-for-machine-learning-audits-on-enwiki/124172/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why is auditing the Wikipedia machine learning models important?","Question",{"text":74,"@type":75},"It helps ensure vandalism-detection systems remain fair and effective. ORES’ edit quality predictions affect whether edits are likely to be reverted, so audit evidence can target unfair or incorrect behavior.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"What makes auditing complex ML systems difficult in this context?",{"text":79,"@type":75},"Auditors must find relevant incorrect predictions, determine whether they reflect a consistent bug-like pattern, and persuade system designers to change the undesired behavior.",{"name":81,"@type":72,"acceptedAnswer":82},"What is ORES-Inspect and what is its main purpose?",{"text":83,"@type":75},"ORES-Inspect is an open-source web tool that enables editors to audit the ORES edit quality model. It functions as a technology probe to study how editors translate intuitions into quantitative evidence about model behavior.","https://schema.org",{"og:url":52,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,112,117,122,127,130,133],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":110,"slug":111},50,"technology",{"id":113,"doc_module":4,"doc_module_name":46,"category_name":114,"show_sort_weight":115,"slug":116},7,"Healthcare",40,"healthcare",{"id":118,"doc_module":4,"doc_module_name":46,"category_name":119,"show_sort_weight":120,"slug":121},8,"Research & Report",30,"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":29,"doc_module":4,"doc_module_name":46,"category_name":131,"show_sort_weight":29,"slug":132},"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":46,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]