[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-86521-en":3,"doc-seo-86521-105":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":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":13,"seo_description":14,"update_tm":28,"read_time":29},86521,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Normative Alignment of Recommender Systems via Internal Label Shift","Recommender systems optimized only for engagement often fall short of broader normative objectives like fairness, diversity, and editorial values. NAILS (Normative Alignment of recommender systems via Internal Label Shift) is a simple, scalable approach that aligns recommendation outputs with target distributions over item attributes such as categories. NAILS adjusts the user-conditional item distribution to induce a specified marginal attribute distribution, using existing user–item preferences without retraining. Framed as an internal label-shift process in a hierarchical classification setting, it improves attribute-level alignment while minimally affecting engagement. ","Normative Alignment of Recommender Systems via Internal  \nLabel Shift  \nJohannes Kruse∗ [johannes.kruse@jppol.dk](johannes.kruse@jppol.dk)[ ](johannes.kruse@jppol.dk)[JP/Politikens Media Group](JP/Politikens Media Group)[ ](JP/Politikens Media Group)Copenhagen, Denmark  \nRyotaro Shimizu‡ [r2shimizu@ucsd.edu](r2shimizu@ucsd.edu)[ ](r2shimizu@ucsd.edu)University of California San Diego La Jolla, California, USA  \nKasper Lindskow† [kasper.lindskow@jppol.dk](kasper.lindskow@jppol.dk)[ ](kasper.lindskow@jppol.dk)[JP/Politikens Media Group](JP/Politikens Media Group)[ ](JP/Politikens Media Group)Copenhagen, Denmark  \nJulian McAuley  \n[jmcauley@eng.ucsd.edu](jmcauley@eng.ucsd.edu)[ ](jmcauley@eng.ucsd.edu)University of California San Diego La Jolla, California, USA  \nMichael Riis Andersen  \n[miri@dtu.dk](miri@dtu.dk)  \nTechnical University of Denmark Kongens Lyngby, Denmark  \nPierre-Alexandre Mattei  \n[pierre-alexandre.mattei@inria.fr](pierre-alexandre.mattei@inria.fr)[ ](pierre-alexandre.mattei@inria.fr)Inria, Université Côte d’Azur Nice, France  \nJes Frellsen§ [jefr@dtu.dk](jefr@dtu.dk)  \nTechnical University of Denmark Kongens Lyngby, Denmark  \narXiv :2607 . 109 15v 1 [ cs .IR] 12 Jul 2026  \nAbstract  \nRecommender systems optimized solely for user engagement often fail to meet broader normative objectives such as fairness, diversity, or editorial values. We introduce NAILS (Normative Alignment of recommender systems via Internal Label Shift), a simple and scalable method for aligning recommendation outputs with target distributions over item-level attributes (e.g., categories) . NAILS modifies the user-conditional item distribution to induce a specified marginal distribution over attributes, leveraging existing user–item preferences without retraining the model. To achieve this, we recast the problem as a form of label shift applied internally within a hierarchical classification framework. Adopting a stakeholdercentric perspective, NAILS enables alignment with global normative goals. Empirically, we show that NAILS consistently improves attribute-level alignment with minimal impact on user engagement, providing a practical mechanism for value-driven recommendation. Our code is available at [https://github.com/johanneskruse/nails](https://github.com/johanneskruse/nails).  \nCCS Concepts  \n• Information systems → Recommender systems.  \nKeywords  \nRecommender Systems; Aligned Recommendation; Normative Design; Relevance Prioritized Reranking  \n1 Introduction  \nRecommender systems play a central role in content delivery across a wide range of domains, including news, entertainment, and ecommerce [4, 9, 14, 24, 25] . While these systems are typically optimized for user engagement, such optimization may come atthe expense of alignment with broader organizational or societal  \n∗ Also with Technical University of Denmark.†Also with Copenhagen Business School.‡Also with ZOZO Research.  \n§ Also with Pioneer Centre for Artificial Intelligence.  \ngoals [7, 31] . In domains like news, for example, purely behaviordriven personalization can reinforce filter bubbles, reduce content diversity, and underrepresent important but less popular content [20, 30] . This growing concern has motivated research into the normative design of recommender systems—designing recommender systems that are not only effective but also align with normative objectives, such as fairness, diversity, or editorial values [6, 11, 16, 18, 28, 29] .  \nIn this paper, we propose Normative Alignment of recommender systems via Internal Label Shift (NAILS), a lightweight framework that enables alignment of recommendations toward normative objectives. While much of the research on alignment of recommendation has adopted a user-centric perspective [1, 10, 27], i.e., focusing on calibrating recommendations based on individual user interactions, our work takes a stakeholder-centric view, empowering platforms, such as mission-driven news publishers, to align recommendations with global n","cbCairZEe6kLeTyZ","https://ap.wps.com/l/cbCairZEe6kLeTyZ","pdf",520651,3,1,6,"English","en",105,"# Introduction\n# Related Work","[{\"question\":\"What problem does NAILS address in recommender systems?\",\"answer\":\"It targets the gap between engagement-optimized recommenders and broader normative goals such as fairness, diversity, or editorial values.\"},{\"question\":\"How does NAILS achieve normative alignment without retraining the model?\",\"answer\":\"NAILS modifies the user-conditional item distribution so that a desired marginal distribution over item attributes emerges, reusing existing user–item preferences.\"},{\"question\":\"On what basis is the method formulated, according to the paper?\",\"answer\":\"The approach reframes the task as an internal label-shift applied within a hierarchical classification framework, supporting a stakeholder-centric alignment perspective.\"}]",1784212354,15,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"normative-alignment-of-recommender-systems-via-internal-label-shift","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"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":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/normative-alignment-of-recommender-systems-via-internal-label-shift/86521/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-25","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does NAILS address in recommender systems?","Question",{"text":75,"@type":76},"It targets the gap between engagement-optimized recommenders and broader normative goals such as fairness, diversity, or editorial values.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does NAILS achieve normative alignment without retraining the model?",{"text":80,"@type":76},"NAILS modifies the user-conditional item distribution so that a desired marginal distribution over item attributes emerges, reusing existing user–item preferences.",{"name":82,"@type":73,"acceptedAnswer":83},"On what basis is the method formulated, according to the paper?",{"text":84,"@type":76},"The approach reframes the task as an internal label-shift applied within a hierarchical classification framework, supporting a stakeholder-centric alignment perspective.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,114,119,122,127,130,134],{"id":21,"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":52,"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":22,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},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":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"]