[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117629-en":3,"doc-seo-117629-105":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},117629,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","Synergy of Machine Learning and Automated Reasoning - Dissertation","Synergy of Machine Learning and Automated Reasoning presents research on combining data-driven machine learning with formal methods for automated reasoning. The work studies premise selection strategies for automated theorem proving, including ATPboost with ATP feedback in binary settings and recurrent neural networks for stateful premise selection over large libraries. It further explores neural guidance in connection tableaux and learning quantifier instantiation in SMT, as well as online machine learning techniques for Coq. Experiments and analyses evaluate effectiveness, robustness, and future directions.","Synergy of Machine Learning and Automated Reasoning  \nProefschrift ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen op gezag van de rector magnificus prof. dr. J.M. Sanders, volgens besluit van het college voor promoties in het openbaar te verdedigen op  \ndinsdag 5 december 2023  \nom 10:30 uur precies  \ndoor  \nBartosz Pawel Piotrowski  \ngeboren op 12 juni 1992  \nte Pruszk´ow, Polen  \nPromotor:  \n• Prof. dr. Herman Geuvers Copromotoren:  \n• Dr. Josef Urban (ˇCesk´e vysok´e uˇcen´ı technick´e v Praze, Tsjechi¨e)  \n• Dr. Mikol´aˇs Janota (ˇCesk´e vysok´e uˇcen´ı technick´e v Praze, Tsjechi¨e) Manuscriptcommissie:  \n• Prof. dr. Tom Heskes  \n• Prof. dr. Stephan Schulz (Duale Hochschule Baden-W¨urttemberg Stuttgart, Duitsland)  \n• Dr. Konstantin Korovin (The University of Manchester, Verenigd Koninkrijk)  \n• Prof. dr. Jasmin Blanchette (Ludwig-Maximilians-Universit¨at M¨unchen, Duitsland)  \n• Prof. dr. Mateja Jamnik (University of Cambridge, Verenigd Koninkrijk)  \nSynergy of Machine Learning and Automated Reasoning  \nDissertation to obtain the degree of doctor from Radboud University Nijmegen on the authority of the Rector Magnificus prof. dr. J.M. Sanders, according to the decision of the Doctorate Board to be defended in public on  \nTuesday, December 5, 2023  \nat 10:30 am  \nby  \nBartosz Pawel Piotrowski  \nborn on June 12, 1992  \nin Pruszk´ow, Poland  \nSupervisor:  \n• Prof. dr. Herman Geuvers Co-Supervisors:  \n• Dr. Josef Urban (Czech Technical University in Prague, Czech Republic)  \n• Dr. Mikol´aˇs Janota (Czech Technical University in Prague, Czech Republic) Manuscript Committee:  \n• Prof. dr. Tom Heskes  \n• Prof. dr. Stephan Schulz (Baden-W¨urttemberg Cooperative State University, Germany)  \n• Dr. Konstantin Korovin (The University of Manchester, United Kingdom)  \n• Prof. dr. Jasmin Blanchette (Ludwig Maximilian University of Munich, Germany)  \n• Prof. dr. Mateja Jamnik (University of Cambridge, United Kingdom)  \nFor it is unworthy of excellent men to lose hours like slaves in the labor of calculation which could be safely relegated to anyone else if the machine were used.  \n— Gottfried Leibniz  \nWe may hope that machines will eventually compete with men in all purely intellectual fields.  \n—Alan Turing  \nContents  \n1 Introduction 1  \n1.1 The quest for mechanized reasoning ................ 1  \n1.1.1 Automated theorem proving ................. 4  \n1.1.2 Interactive theorem proving ................. 6  \n1.2 The promise of learning from examples ............... 9  \n1.3 Automated reasoning meets machine learning ........... 11  \n1.3.1 Incorporating data-driven paradigm into formal tools ... 12  \n1.3.2 Challenges of making machine learning enhancements .. 14  \n1.3.3 Can machine learning methods reason on their own? ... 16  \n1.4 Thesis outline ............................. 18  \n1.4.1 Structure of the thesis .................... 18  \n1.4.2 Main research questions ................... 20  \n2 ATPboost: Learning premise selection in binary setting with ATP feedback 23  \n2.1 Introduction: Machine learning for premise selection ....... 24  \n2.1.1 Premise selection in binary setting with multiple proofs . 25  \n2.2 ATPboost: Setting, algorithms and components .......... 26  \n2.2.1 Algorithms .......................... 26  \n2.2.2 Components ......................... 27  \n2.3 Evaluation ............................... 30  \n2.3.1 Parameter tuning ....................... 30  \n2.3.2 Incremental feedback loop with train/test split ...... 32  \n2.3.3 Incremental feedback loop with no initial proofs ..... 32  \n2.4 Conclusions and future work .................... 34  \n3 Stateful premise selection by recurrent neural networks 37  \n3.1 Introduction: Premise selection over large libraries ........ 37  \n3.2 Premise selection and neural machine translation ......... 38  \n3.3 Data, their representation and augmentation ........... 39  \n3.3.1 Initial data for training RNNs ................ 40  \n3.3.2 Representation of the statements","cbCaiu0yKEb1fD5m","https://ap.wps.com/l/cbCaiu0yKEb1fD5m","pdf",3374880,1,180,"English","en",105,"# Introduction\n## The quest for mechanized reasoning\n## Automated theorem proving\n## Interactive theorem proving\n## The promise of learning from examples\n## Automated reasoning meets machine learning\n## Incorporating data-driven paradigm into formal tools\n## Challenges of making machine learning enhancements\n## Can machine learning methods reason on their own?\n## Thesis outline\n## Structure of the thesis\n## Main research questions\n# ATPboost: Learning premise selection in binary setting with ATP feedback\n## Introduction: Machine learning for premise selection\n## Premise selection in binary setting with multiple proofs\n## ATPboost: Setting, algorithms and components\n## Algorithms\n## Components\n## Evaluation\n## Parameter tuning\n## Incremental feedback loop with train/test split\n## Incremental feedback loop with no initial proofs\n## Conclusions and future work\n# Stateful premise selection by recurrent neural networks\n## Introduction: Premise selection over large libraries\n## Premise selection and neural machine translation\n## Data, their representation and augmentation\n## Initial data for training RNNs\n## Representation of the statements\n## Ordering of the premises\n## Augmentation with subproof data\n## Oversampling rare examples\n## Experimental evaluation\n## Results and discussion\n## Source and target combinations\n## Augmentation with subproof data and oversampling\n## Subproofs as standalone data set\n## Examples of predictions from RNN\n## Conclusions and future work\n# Guiding inferences in connection tableau by recurrent neural networks\n## Introduction\n## A data set for connection-style internal guidance\n## Neural modelling and evaluation metric\n## Results\n## Conjecturing new literals\n## Conclusions and future work\n# Towards learning quantifier instantiation in SMT\n## Introduction\n## Background\n## Enumerative instantiation\n## Learning ordering of terms\n## Featurization\n## Experimental evaluation\n## Experimental setting\n## Data for evaluation\n## Results and discussion\n## Related work\n## Conclusions and future work\n# Online machine learning techniques for Coq\n## Introduction\n## Contributions\n## Tactic and proof state representation","[{\"question\":\"What is the main theme of the dissertation?\",\"answer\":\"The dissertation focuses on combining machine learning with automated reasoning, improving formal theorem proving and related inference tasks using data-driven methods.\"},{\"question\":\"How does ATPboost improve premise selection?\",\"answer\":\"ATPboost learns premise selection in a binary setting using ATP feedback, with evaluation including parameter tuning and incremental feedback loop variants.\"},{\"question\":\"Which reasoning environments beyond ATP are addressed?\",\"answer\":\"The dissertation extends learning-based guidance to connection tableau methods, SMT quantifier instantiation, and online machine learning techniques for Coq.\"}]","Synergy of Machine Learning and Automated Reasoning - Dissertation | PDF",1785677439,454,{"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},"synergy-of-machine-learning-and-automated-reasoning-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/synergy-of-machine-learning-and-automated-reasoning-dissertation/117629/",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},"What is the main theme of the dissertation?","Question",{"text":75,"@type":76},"The dissertation focuses on combining machine learning with automated reasoning, improving formal theorem proving and related inference tasks using data-driven methods.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does ATPboost improve premise selection?",{"text":80,"@type":76},"ATPboost learns premise selection in a binary setting using ATP feedback, with evaluation including parameter tuning and incremental feedback loop variants.",{"name":82,"@type":73,"acceptedAnswer":83},"Which reasoning environments beyond ATP are addressed?",{"text":84,"@type":76},"The dissertation extends learning-based guidance to connection tableau methods, SMT quantifier instantiation, and online machine learning techniques for Coq.","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"]