[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-291109-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-291109-en":131},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","re-evaluating-the-machine-learning-pipeline-to-improve-fairness-and-reliability","Re-Evaluating the Machine Learning Pipeline to Improve Fairness and Reliability","","Across the world, an increasing number of decisions are at least partially automated using machine learning (ML) or artificial intelligence (AI) models in algorithmic decision-making (ADM). The introduction of ADM systems has transformed many domains, but also raised ethical concerns. Bias, discrimination, and insufficient transparency or accountability can lead to unfair outcomes. Building such systems requires a pipeline of data creation, processing, model selection, training, evaluation, and deployment.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/re-evaluating-the-machine-learning-pipeline-to-improve-fairness-and-reliability/291109/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/re-evaluating-the-machine-learning-pipeline-to-improve-fairness-and-reliability/291109.png","ImageObject",300,407,{"name":42,"@type":43},"Theodora","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-20","2026-09-17",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",11,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What problem does the thesis focus on?","Question",{"text":63,"@type":64},"The thesis re-evaluates the traditional machine learning pipeline to address fairness and reliability concerns in algorithmic decision-making.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"Which stages are part of the machine learning pipeline discussed?",{"text":68,"@type":64},"The pipeline includes data creation or choice, (pre-)processing, model selection and architecture, training, evaluation, and deployment.",{"name":70,"@type":61,"acceptedAnswer":71},"How does the thesis conceptualize the impact of pipeline decisions?",{"text":72,"@type":64},"It views ad hoc, interacting choices as a “garden of forking paths,” and analyzes possible outcomes as a multiverse explored via multiverse analysis.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},291109,1789643475,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,115,120,123,127],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":112,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":113,"slug":114},8,30,"research-report",{"id":116,"doc_module":4,"doc_module_name":25,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":25,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":25,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":25,"category_name":129,"show_sort_weight":97,"slug":130},19,"General","general",{"code":4,"msg":82,"data":132},{"doc_id":79,"user_id":133,"nickname":42,"user_avatar":134,"doc_module":4,"category_id":112,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":140,"language":141,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":12,"update_tm":80,"read_time":145,"preview_page_count":4},687197207919,"https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552","Jan Simson  \nRe-Evaluating the Machine Learning Pipeline to Improve Fairness and Reliability  \nDissertation an der Fakultt fr Mathematik, Informatik und Statistik der Ludwig-Maximilians-Universit t M nchen  \nEingereicht am 3.11.2025  \nJan Simson  \nRe-Evaluating the Machine Learning Pipeline to Improve Fairness and Reliability  \nDissertation an der Fakult t fr Mathematik, Informatik und Statistik der Ludwig-Maximilians-Universit t M nchen  \nEingereicht am 3.11.2025  \nErster Berichterstatter: Prof. Dr. Christoph Kern Zweiter Berichterstatter: Prof. Dr . Frauke Kreuter Dritter Berichterstatter: Prof. Dr. Roger Peng  \nTag der Disputation: 20.01.2026  \nAcknowledgments  \nA huge thank you to all the amazing people who have made the last three years not only possible, but actually incredibly fun. It was a real privilege getting to know you all, and every single one of you made a positive difference along this journey.  \n■ First and foremost, I want to thank my advisors, Christoph and Frauke, for their continuous support, trust, and guidance. I especially appreciate the flexibility and freedom, alongside gentle nudges, that you provided, allowing my research to be shaped by my interests while staying on track. Thank you for the opportunity to do a PhD under your supervision!  \n■ A big thank you to Roger Peng for acting as the third reviewer on this thesis, as well as Joseph Sakshaug and Helmut K¨uchenhoff for forming the examination committee for the disputation together.  \n■ Thank you to the wonderful collaborators I got to work with over the past three years: Alessandro, Sam, Fiona, Florian, Cosima, Malte, and Olga. I appreciate the constructive discussions, valuable contributions, and great company you have all provided for these works. A special thank you to Sam for his advice and guidance in navigating the academic landscape. I hope and look forward to future collaborations with all of you!  \n■ Thank you to Patrick, Yegi, and Cynthia for reading earlier drafts of this thesis and the valuable feedback they provided to improve it.  \n■ Thank you to the past and present members of SODA. Thank you to Ailin, Andreas, Anna, Bolei, Caro, Clara, Felicitas, Felix, Fiona, Julia, Leo, Lisa, Luis, Malte, Marcus, Markus, Olga, Patrick, Ruben, Sarah, Sofia, Unai, and Wiebke. I appreciate the mensa visits, coffee chats, Pf¨alzer celebrations, board game evenings, and especially the trips and visits together with many of you. Special thanks to Jacob for the warm welcome and rising to YouTube fame together and to Leah, without whom I would’ve never known about this opportunity. Iam grateful to get to call so many of you friends rather than colleagues.  \n■ Thank you to Pia, Yvonne, Christina, Brigitte, and Denise for their continued help with the hardest part of research. You make it almost seem easy to navigate German university bureaucracy, and I enjoy every opportunity for a small chat on the 1st floor.  \n■ I would also like to thank all the other lovely people at the Institute of Statistics here at LMU. Everyone has been incredibly welcoming, and it has been a joy to get to know and befriend many of you. It’s truly an extraordinary institute. A special thank you goes out to Yegi, who has been the best of friends from the very beginning and to whom I owe more things than can be listed here – your place in Munich is empty. Special thanks also to Alex and Martje for more than a card deck’s worth of fond memories and Yin for being extra welcoming from day one.  \n■ Thank you also to the newfound friends outside of the institute, who have played a significant role in making Munich feel like home. Thank you for the movie nights, Padel sessions, concerts, and, most importantly, of course, all the hikes.  \n■ Last, I’d like to thank my parents and brother for being the solid foundation that has always made it easy to explore.  \nAbstract  \nAcross the world, an increasing number of decisions are, at least partially, automated using machine learning (ML) or ","cbCaibtUwka9QzDW","https://ap.wps.com/l/cbCaibtUwka9QzDW","pdf",6772122,180,"English","# Abstract\n## Algorithmic decision-making and ethical challenges\n## Pipeline construction and decision points\n## Multiverse analysis across pipeline choices\n## Thesis contribution and re-evaluation of traditional pipelines","[{\"question\":\"What problem does the thesis focus on?\",\"answer\":\"The thesis re-evaluates the traditional machine learning pipeline to address fairness and reliability concerns in algorithmic decision-making.\"},{\"question\":\"Which stages are part of the machine learning pipeline discussed?\",\"answer\":\"The pipeline includes data creation or choice, (pre-)processing, model selection and architecture, training, evaluation, and deployment.\"},{\"question\":\"How does the thesis conceptualize the impact of pipeline decisions?\",\"answer\":\"It views ad hoc, interacting choices as a “garden of forking paths,” and analyzes possible outcomes as a multiverse explored via multiverse analysis.\"}]","Re-Evaluating the Machine Learning Pipeline to Improve Fairness and Reliability | PDF",454]