[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118038-en":3,"doc-seo-118038-105":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},118038,549758252649,"Ivy","https://ap-avatar.wpscdn.com/avatar/8000253669c5317157?_k=1778319167496531819",8,"Research & Report","AUTOMATED MACHINE LEARNING WITH CONSTRAINTS AND IMPERFECT DATA - Doctoral Dissertation","Automated machine learning (AutoML) accelerates the discovery of high-performance models by reducing manual tuning, yet it faces real-world issues: datasets can contain wrong labels, limited data size, and imbalanced label distributions. Label noise disrupts training and undermines representative evaluation, while imbalance skews search feedback. Additional constraints—model size, fairness, and robustness—also complicate the AutoML pipeline, and insufficient computing resources restrict search space and can create mismatches between search and evaluation architectures. This dissertation develops constraint-aware methods for robust learning, imbalance-aware search, defect generation, adaptive early stopping, and efficient model parallelism across multiple GPUs.","AUTOMATED MACHINE LEARNING WITH CONSTRAINTS AND IMPERFECT DATA  \nA Dissertation  \nby  \nYI-WEI CHEN  \nSubmitted to the Graduate and Professional School of  \nTexas A&M University  \nin partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY  \nChair of Committee, Xia Hu Committee Members, P.R. Kumar  \nEun Jung Kim  \nZhangyang Wang  \nHead of Department, Scott Schaefer  \nDecember 2022  \nMajor Subject: Computer Science  \nCopyright 2022 Yi-Wei Chen  \nABSTRACT  \nMachine learning has succeeded in real-world applications from image classification, speech recognition, to beating human champion in Go games. To accelerate the development of different applications, automated machine learning (AutoML) has been proposed to discover high-performance machine learning models automatically. It could release the burden of data scientists from the multifarious manual tuning process. However, dataset does not always have correct labels and sufficient data size. Wrong labels disrupt the training procedure and could not provide representative evaluation performance for AutoML. Imbalanced label distribution further skews search feedback for AutoML. Furthermore, additional performance constraints, such as model size, fairness, and robustness complicates the AutoML flow. When computing resources are insufficient, small search space constrains the flexibility to search neural networks, which causes inconsistent architectures used in search and evaluation stages. In this dissertation, I advanced AutoML from aspects of imperfect data and constraints. A new robust loss function is integrated with search algorithm for label noise. I design a simple but effective search space for imbalanced defect datasets. The defect generator can alleviate imbalanced distributions. I also proposed constraint-aware early stopping for AutoML with adaptive constraint evaluation intervals. An efficient model parallelism for AutoML is proposed to extend search spaces in multiple GPUs with limited memory size. My research of automated machine learning enables scientists to obtain off-to-shelf models on various data formats, as well as customizes models for different computing resources, model size requirements, and miscellaneous performance constraints. It broadly impacts image classification, constrained  \nAutoML, and defect detection.  \nDEDICATION  \nTo Meng-Hua Guo.  \nACKNOWLEDGMENTS  \nI would like to thank Dr. Xia Hu for strong support of my research. He encouraged me to complete the AutoML survey in my early research stage, making me understand the field in width and depth. He is open-minded to discuss a variety of research topics, creating an active and productive working atmosphere. I sincerely appreciate his steady supports for my life, research, internship, and my future career. His generous and industrious attitude inspires me to become abetter researcher. I would also like to thank the rest of my dissertation committee, Dr. Zhangyang Wang, Dr. Eun Jung Kim, and Dr. P.R. Kumar for their advice. I am grateful to Dr. Chi Wang, Dr. Amin Saied, and Dr. Rui Zhuang, who gave me inspiring research discussions during my research project at Microsoft. I would like to thank Dr. Chu Wang and Ms. Guangyu Zhang for their hlep and advice on my remote internship at Amazon. Many thanks to my outstanding lab colleagues at DATA Lab at Texas A&M University and Rice University. I enjoyed all research discussions and brainstorming. Special thanks to all the reviewers to my papers, the Texas A&M University, and all the funding agencies, including National Science Foundation and Defense Advanced Research Projects Agency.  \nFinally, I would love to express my sincere gratitude to my family, my mother, Qiu-Li Huang, my father, Sen-Xian Chen, mother-in-law, Su-Chin Jiang, father-in-law, Shui-Quan Guo, for their endless encouragement and love. I epxress the greatest thankfulness to my beloved wife, Meng-Hua  \nGuo. All my achievements belong to her.  \nCONTRIBUTORS AND FUNDING S","cbCaibh1ALXCWXJb","https://ap.wps.com/l/cbCaibh1ALXCWXJb","pdf",3285738,1,102,"English","en",105,"# Abstract\n# Dedication\n# Acknowledgments\n# Contributors and Funding Sources\n# Table of Contents\n# List of Figures\n# List of Tables\n# 1. Introduction\n## 1.1 Motivation and Challenges","[{\"question\":\"How does label noise affect automated machine learning in this dissertation?\",\"answer\":\"Wrong labels disrupt the training procedure and prevent AutoML from producing representative evaluation performance. The dissertation integrates a new robust loss function with the search algorithm to handle label noise.\"},{\"question\":\"What problems arise from imbalanced label distributions, and how is imbalance addressed?\",\"answer\":\"Imbalanced label distributions skew search feedback for AutoML. The work designs an effective search space for imbalanced defect datasets and introduces a defect generator to alleviate imbalance.\"},{\"question\":\"How are additional performance constraints handled during AutoML?\",\"answer\":\"Constraints such as model size, fairness, and robustness complicate the AutoML flow. The dissertation proposes constraint-aware early stopping with adaptive constraint evaluation intervals, and also studies extending search spaces under limited GPU memory via model parallelism.\"}]","AUTOMATED MACHINE LEARNING WITH CONSTRAINTS AND IMPERFECT DATA - Doctoral Dissertation | PDF",1785680949,257,{"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},"automated-machine-learning-with-constraints-and-imperfect-data-doctoral-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/automated-machine-learning-with-constraints-and-imperfect-data-doctoral-dissertation/118038/",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},"How does label noise affect automated machine learning in this dissertation?","Question",{"text":75,"@type":76},"Wrong labels disrupt the training procedure and prevent AutoML from producing representative evaluation performance. The dissertation integrates a new robust loss function with the search algorithm to handle label noise.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What problems arise from imbalanced label distributions, and how is imbalance addressed?",{"text":80,"@type":76},"Imbalanced label distributions skew search feedback for AutoML. The work designs an effective search space for imbalanced defect datasets and introduces a defect generator to alleviate imbalance.",{"name":82,"@type":73,"acceptedAnswer":83},"How are additional performance constraints handled during AutoML?",{"text":84,"@type":76},"Constraints such as model size, fairness, and robustness complicate the AutoML flow. The dissertation proposes constraint-aware early stopping with adaptive constraint evaluation intervals, and also studies extending search spaces under limited GPU memory via model parallelism.","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"]