[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117511-en":3,"doc-seo-117511-105":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},117511,1099514067438,"River Wang","https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542",8,"Research & Report","Steering Machine Learning Ecosystems of Interacting Agents - Dissertation","Steering Machine Learning Ecosystems of Interacting Agents studies how deployed ML systems, including large language models and recommender systems, generate unintended ecosystem-level effects when interacting with humans and companies. It argues that standard single-agent evaluation misses outcomes at the society, market, and algorithm levels. The thesis traces these outcomes to agent incentives and to the training pipeline. It develops tools for assessing AI policy under provider competition, analyzes recommendation-driven creator incentives and generative participation patterns, and proposes incentive-aware evaluation metrics and algorithms for repeated human–model interaction.","UC Berkeley  \nUC Berkeley Electronic Theses and Dissertations  \nTitle  \nSteering Machine Learning Ecosystems of Interacting Agents  \nPermalink  \n[https://escholarship.org/uc/item/9h2365bx](https://escholarship.org/uc/item/9h2365bx)  \nISBN  \n9798288866081  \nAuthor  \nJagadeesan, Meena  \nPublication Date  \n2025-05-24  \nPeer reviewed  \n[eScholarship.org](eScholarship.org) Powered by the California Digital Library  \nUniversity of California  \nSteering Machine Learning Ecosystems of Interacting Agents  \nBy  \nMeena Jagadeesan  \nA dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy in  \nComputer Science  \nin the  \nGraduate Division  \nof the  \nUniversity of California, Berkeley  \nCommittee in charge:  \nProfessor Michael I. Jordan, Co-Chair Assistant Professor Jacob Steinhardt, Co-Chair Assistant Professor Nika Haghtalab Associate Professor Anca D. Dragan Professor Federico Echenique  \nSpring 2025  \nSteering Machine Learning Ecosystems of Interacting Agents  \nCopyright 2025  \nBy  \nMeena Jagadeesan  \n1  \nAbstract  \nSteering Machine Learning Ecosystems of Interacting Agents  \nBy  \nMeena Jagadeesan  \nDoctor of Philosophy in Computer Science  \nUniversity of California, Berkeley  \nProfessor Michael I. Jordan, Co-Chair  \nAssistant Professor Jacob Steinhardt, Co-Chair  \nWhen machine learning models such as large language models (LLMs) and recommender systems are deployed into human-facing applications, these models interact with humans, companies, and other models within a broader ecosystem. However, the resulting multiagent interactions often induce unintended ecosystem-level outcomes, including clickbait in classical content recommendation ecosystems, and more recently, safety violations and market concentration in nascent LLM ecosystems. The core issue is that ML models are classically analyzed as a single agent operating in isolation, so standard evaluation approaches in machine learning fail to capture ecosystem-level outcomes at the society-level, market-level, and algorithm-level.  \nThis thesis investigates how to characterize and steer ecosystem-level outcomes, focusing on LLM ecosystems and content recommendation ecosystems. To tackle this, we augment the typical algorithmic perspective on machine learning with an economic and statistical perspective. The key idea is to trace ecosystem-level outcomes back to the incentives of interacting agents (i.e., ML models, humans, and companies) and back to the ML pipeline for training models.  \nIn the first part, we investigate how competition between model-providers influences ecosystemlevel performance trends and market outcomes. We demonstrate that scaling trends are fundamentally altered, and we develop technical tools to evaluate proposed AI policy. In the second part, we investigate how ML models deployed in content recommendation ecosystems influence content creation. We characterize how recommendation models shape the content supply via creator incentives, and how generative models shape which types of users produce content. In the third part, we investigate repeated interactions between a human and a ML model. We develop evaluation metrics which account for competing preferences, and design near-optimal incentive-aware algorithms.  \n2  \nMore broadly, this thesis takes a step towards a vision of machine learning ecosystems where the interactions between ML models, humans, and companies are steered towards the desired ecosystem-level outcomes.  \ni  \nTo my family and friends, for their love, support, and laughter.  \nii  \nContents  \nContents ii  \nList of Figures vii  \nList of Tables xiv  \nI Introduction 1  \n1 Overview 2  \n1.1 Our contributions ................................. 3  \nII Model-Provider Competition 6  \n2 Overview 7  \n2.1 Our contributions ................................. 7  \n2.2 Methodological theme .............................. 8  \n3 Developers Fine-tuning a Pretrained Model 9  \n3.1 Introduction ..........................","cbCaiuaF9Nzh1cas","https://ap.wps.com/l/cbCaiuaF9Nzh1cas","pdf",14589112,1,651,"English","en",105,"# I Introduction\n## 1 Overview\n## 1.1 Our contributions\n# II Model-Provider Competition\n## 2 Overview\n## 2.1 Our contributions\n## 2.2 Methodological theme\n## 3 Developers Fine-tuning a Pretrained Model\n## 4 Companies Training Language Models\n## 5 Recommendation Platforms\n## 6 The Power of a Digital Platform\n# III Incentives for Digital Content Creation\n## 7 Overview\n## 7.1 Our contributions\n## 7.2 Methodological theme\n## 8 Specialized vs. Homogenized Content","[{\"question\":\"Why do classic machine learning evaluations fail to capture ecosystem-level outcomes?\",\"answer\":\"ML models are typically analyzed as single agents in isolation, so standard metrics cannot reflect society-, market-, and algorithm-level effects caused by multiagent interactions.\"},{\"question\":\"What is the thesis’s central approach to steering ecosystem outcomes?\",\"answer\":\"It combines an algorithmic view with economic and statistical perspectives, tracing ecosystem effects back to the incentives of interacting agents and to the ML training pipeline.\"},{\"question\":\"How does the thesis handle repeated human–model interactions?\",\"answer\":\"It introduces evaluation metrics that account for competing preferences and designs near-optimal incentive-aware algorithms for these repeated interactions.\"}]","Steering Machine Learning Ecosystems of Interacting Agents - Dissertation | PDF",1785676478,1641,{"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},"steering-machine-learning-ecosystems-of-interacting-agents-dissertation","",{"@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/steering-machine-learning-ecosystems-of-interacting-agents-dissertation/117511/",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 do classic machine learning evaluations fail to capture ecosystem-level outcomes?","Question",{"text":75,"@type":76},"ML models are typically analyzed as single agents in isolation, so standard metrics cannot reflect society-, market-, and algorithm-level effects caused by multiagent interactions.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the thesis’s central approach to steering ecosystem outcomes?",{"text":80,"@type":76},"It combines an algorithmic view with economic and statistical perspectives, tracing ecosystem effects back to the incentives of interacting agents and to the ML training pipeline.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the thesis handle repeated human–model interactions?",{"text":84,"@type":76},"It introduces evaluation metrics that account for competing preferences and designs near-optimal incentive-aware algorithms for these repeated interactions.","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"]