[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127848-en":3,"doc-seo-127848-105":31,"detail-sidebar-cat-0-en-105":92},{"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":28,"seo_description":14,"update_tm":29,"read_time":30},127848,2336474466712,"Maeve","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Improving Formal Reasoning for Software Systems using Machine Learning and Symbolic Approaches - Dissertation","Developing reliable software systems is essential for modern society, yet software failures can cause major economic losses, privacy violations, and safety risks. Formal reasoning strengthens reliability through precise specifications and correctness verification, but scalability remains limited by high computational complexity. This dissertation develops methods and tools that integrate symbolic techniques with machine learning to improve scalability and efficiency across both front-end and back-end layers of formal reasoning. It presents iAlloy for incremental Alloy analysis, NeuroBack for CPU-based GNN-enhanced SAT solving, and SymMC for efficient model counting.","Copyright by  \nWenxi Wang 2024  \n1  \nThe Dissertation Committee for Wenxi Wang certifies that this is the approved version of the following dissertation:  \nImproving Formal Reasoning for Software Systems using Machine Learning and Symbolic Approaches  \nCommittee:  \nSarfraz Khurshid, Supervisor  \nKen McMillan  \nRisto Miikkulainen  \nDavid Soloveichik  \nVijay K. Garg  \nImproving Formal Reasoning for Software Systems using Machine Learning and Symbolic Approaches  \nby  \nWenxi Wang  \nDissertation  \nPresented to the Faculty of the Graduate School of The University of Texas at Austin  \nin Partial Fulfillment  \nof the Requirements  \nfor the Degree of  \nDoctor of Philosophy  \nThe University of Texas at Austin August 2024  \nDedication  \nThis dissertation is dedicated to my grandparents and husband, whose boundless love have been my greatest strengths.  \nAcknowledgments  \nFirst, I want to express my deep gratitude to my supervisor, Sarfraz Khurshid. Sarfraz has been like a father to me throughout my academic journey. His immense support during the early years of my PhD and his openness in allowing me to pursue independent research later on have been invaluable. Thank you, Sarfraz, for always being there whenever I needed help and for advocating for my job.  \nI also want to extend special thanks to my academic uncle, Darko Marinov, and my collaborators, Kenneth McMillan and Risto Miikkulainen. They were guiding lights during my independent research phase. My heartfelt thanks also go to my academic brother, Lingming Zhang, for his helpful advice and for introducing me to many amazing people in our field. Additionally, I am grateful to my academic cousins, Milos Gligoric, August Shi, Wing Lam, and Owolabi Legunsen, for their support during my job search.  \nNext, I want to thank my family. Special thanks to my grandfather, the most important person in my life, who gave me a kind and strong heart and was the first person in my family to believe in my dream of becoming a professor. I would also like to thank my husband, Yang Hu, my best collaborator and friend, for providing immense technical and emotional support in helping me realize my dream step by step. Thanks to my grandmother, my parents, my sister Wenchen Wang, my aunt, and my cousin Wen Jiang for making me feel so loved ever since I was born. Their love has supported me through the hard days of my PhD journey.  \nI am also grateful to my friends, roommates, classmates, and officemates from my bachelor's and master's studies who gave me immense love and attention, making me feel special and cherished. I also want to extend special thanks to my lovely friends and officemates in Austin: Yuki Liu, Jiayi Yang, Yating Wu, Zhiqiang Zang, Pengyu Nie, Chengpeng Li, and Marko Vasic. Their companionship and friendship have been  \nessential for my happiness throughout my PhD life. Without them, I would not have enjoyed such a wonderful life in Austin.  \nLastly, I am thankful for the financial support that made my research possible, including the US National Science Foundation under Grants Nos. CCF-1566363, CCF-1652517, CCF-1704790, and CCF-1718903, and a grant from the Army Research Office accomplished under Cooperative Agreement Number W911NF-19-2-0333 .  \nAbstract  \nImproving Formal Reasoning for Software Systems using Machine Learning and Symbolic Approaches  \nWenxi Wang, PhD  \nThe University of Texas at Austin, 2024  \nSUPERVISOR: Sarfraz Khurshid  \nDeveloping reliable software systems is essential as modern society heavily relies on software for critical functions and everyday operations. Software failures can result in severe economic losses, privacy violations, and safety risks. Formal reasoning plays a vital role in enhancing software reliability by providing precise specifications, verifying correctness, and systematically eliminating bugs. However, it often struggles with scalability due to high computational complexity.  \nThis dissertation focuses on creating methods and tools in formal reasoning to","cbCaioZnVcw101Un","https://ap.wps.com/l/cbCaioZnVcw101Un","pdf",3162084,2,1,108,"English","en",105,"# Abstract\n## Front-end: iAlloy for Alloy specification analysis\n## Back-end: NeuroBack for SAT solving with graph neural networks\n## Combined front-end and back-end: SymMC for model counting","[{\"question\":\"Why does the dissertation focus on improving formal reasoning for software systems?\",\"answer\":\"Software failures can lead to economic losses, privacy violations, and safety risks. Formal reasoning increases reliability through precise specifications and correctness verification, but it often struggles to scale due to computational complexity.\"},{\"question\":\"What is iAlloy and what problem does it address?\",\"answer\":\"iAlloy is an incremental analysis method for Alloy specifications. It reduces development costs when users make incremental changes and rerun the analyzer, cutting SAT solver calls by over 50% on average and reaching up to 7x speedups.\"},{\"question\":\"How does NeuroBack improve SAT solving and where is machine learning used?\",\"answer\":\"NeuroBack uses graph neural networks to predict key variable phases before SAT solving begins and performs offline model inference on the CPU. Trained with the DataBack dataset, it improves Kissat performance by solving up to 7.4% more problems in recent SAT competitions.\"}]","Improving Formal Reasoning for Software Systems using Machine Learning and Symbolic Approaches - Dissertation | PDF",1785942337,272,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"improving-formal-reasoning-for-software-systems-using-machine-learning-and-symbolic-approaches-dissertation","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/improving-formal-reasoning-for-software-systems-using-machine-learning-and-symbolic-approaches-dissertation/127848/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-23","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why does the dissertation focus on improving formal reasoning for software systems?","Question",{"text":76,"@type":77},"Software failures can lead to economic losses, privacy violations, and safety risks. Formal reasoning increases reliability through precise specifications and correctness verification, but it often struggles to scale due to computational complexity.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What is iAlloy and what problem does it address?",{"text":81,"@type":77},"iAlloy is an incremental analysis method for Alloy specifications. It reduces development costs when users make incremental changes and rerun the analyzer, cutting SAT solver calls by over 50% on average and reaching up to 7x speedups.",{"name":83,"@type":74,"acceptedAnswer":84},"How does NeuroBack improve SAT solving and where is machine learning used?",{"text":85,"@type":77},"NeuroBack uses graph neural networks to predict key variable phases before SAT solving begins and performs offline model inference on the CPU. Trained with the DataBack dataset, it improves Kissat performance by solving up to 7.4% more problems in recent SAT competitions.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]