[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122569-en":3,"doc-seo-122569-105":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},122569,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","DOTTORATO DI RICERCA IN COMPUTER SCIENCE AND ENGINEERING - MACHINE LEARNING FOR SOFTWARE ENGINEERING - Thesis overview","Open-source repositories are rapidly expanding, creating both opportunities and significant technical challenges for Machine Learning for Software Engineering (ML4SE). Existing approaches are hindered by code transformations such as reformatting or minification, by the absence of standardized benchmarks for repository recommendation, and by the need to operate at archive scale covering billions of files. This thesis introduces transformation-resilient stylometry based on CST paths, performs a large-scale systematic mapping of repository-recommendation research, and proposes DRAGON, a multi-label sentence-pair BERT classifier designed for archive-scale data.","DOTTORATO DI RICERCA IN  \nCOMPUTER SCIENCE AND ENGINEERING  \nCiclo 37  \nSettore Concorsuale: 01/B1-INFORMATICA  \nSettore Scientifico Disciplinare: INF/01-INFORMATICA  \nMACHINE LEARNING FOR SOFTWARE ENGINEERING  \nPresentata da: Stefano Balla  \nCoordinatore Dottorato  \nIlaria Bartolini  \nSupervisore  \nMaurizio Gabbrielli  \nCo-supervisore  \nStefano Zacchiroli  \nAbstract  \nThe explosive growth of open-source repositories creates opportunities and challenges for Machine Learning for Software Engineering (ML4SE) . Current methods struggle with: (i) code frequently reformatted or minified, obscuring stylistic signals; (ii) the lack of standardised, common benchmarks for repository recommendation; and (iii) the need to scale to billions of files held in archives such as Software Heritage.  \nObjectives. This thesis aims to (1) develop an authorship-attribution technique that remains reliable after common code transformations,(2) conduct the first large-scale systematic mapping of repository-recommendation research, and (3) design a multi-label classifier that operates at archive scale.  \nMethods. A language-agnostic stylometric representation based on Concrete Syntax Tree (CST) path-contexts is introduced, contrasted with traditional Abstract Syntax Trees (ASTs) . A systematic mapping study screens over 1 700 papers and distils 43 primary studies, revealing gaps in benchmark standardisation and scalability. To address these gaps, DRAGON is proposed, a sentence-pair BERT model with focal loss and adaptive thresholding, trained on 825 k repositories and 239 GitRanking topics.  \nResults. On untransformed code, CST-based stylometry lifts top-1 author-recognition accuracy from 51 % to 68 %, a 17 % absolute gain over AST baselines. After formatting or minification, recognition falls for both representations, yet CST still leads, delineating the limited privacy these transformations afford. DRAGON raises F 1 @5 by 11 % over the best prior work, in relative terms, and maintains this performance even when 34 % of projects lack a README file. All datasets, model checkpoints, and evaluation scripts are released under permissive open licences.  \nContributions. (i) A transformation-resilient stylometry pipeline; (ii) the largest systematic map of repository-recommendation research to date; (iii) the first repository classifier evaluated on Software-Heritage-scale data; and (iv) practical guidelines for high-fidelity representation, scale-aware engineering, and ethically responsible deployment.  \nImpact. The findings enable accurate topic tagging, strengthen forensic analysis, and guide the design of ML4SE systems able to keep pace with the expansion of open-source software ecosystems.  \nKeywords: ML4SE, authorship attribution, topic tagging, repository classification, software archives, privacy.  \n2  \nContents  \n1 Introduction 7  \n1.1 Research Objectives ........................ 8  \n1.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . 8  \n1.3 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 9  \n1.4 Overview of the Thesis . . . . . . . . . . . . . . . . . . . . . . 10  \n2 Background 11  \n2.1 Stylometry ............................. 12  \n2.2 Code representation   12  \n2.2.1 Abstract Syntax Tree ................... 12  \n2.2.2 Concrete Syntax Tree ................... 13  \n2.2.3 Paths and Path Contexts in CSTs . . . . . . . . . . . . 14  \n2.3 Code formatting   16  \n2.4 Code minification . . . . . . . . . . . . . . . . . . . . . . . . . 17  \n2.5 Software Archives and Storage Solutions   18  \n2.5. 1 GitHub . . . . . . . . . . . . . . . . . . . . . . . . . . 19  \n2.5.2 Software Heritage ..................... 20  \n2.6 Software Repository Taxonomies ................. 21  \n2.6.1 Design Dimensions of Repository Taxonomies ..... 22  \n2.6.2 Rationale for Choosing GitRanking ........... 24  \n2.7 Models in Machine Learning for Software Engineering ..... 25  \n2.7. 1 Code2Vec . . . . . . . . . . . . . . . . . . . . . . . . . 25  \n","cbCaigwQRVNS7iqa","https://ap.wps.com/l/cbCaigwQRVNS7iqa","pdf",1907599,1,141,"English","en",105,"# Introduction\n## Research Objectives\n## Problem Statement\n## Contributions\n## Overview of the Thesis\n# Background\n## Stylometry\n## Code representation\n## Code formatting\n## Code minification\n## Software Archives and Storage Solutions\n## Software Repository Taxonomies\n## Models in Machine Learning for Software Engineering\n# Code Stylometry vs Formatting and Minification\n## Introduction\n## Related Work\n## Methodology\n## Results\n## Discussion\n## Conclusion\n# Software Repositories Classification: A Systematic Mapping Study\n## Introduction\n## Methodology\n## Results","[{\"question\":\"What challenges motivate this ML4SE thesis?\",\"answer\":\"The thesis targets three core issues: code is often reformatted/minified, benchmarks for repository recommendation are not standardized, and models must scale to archive-scale datasets such as Software Heritage.\"},{\"question\":\"How does the work represent code to improve robustness to transformations?\",\"answer\":\"It proposes a language-agnostic stylometric representation using Concrete Syntax Tree (CST) path-contexts, contrasted with traditional Abstract Syntax Tree (AST) approaches.\"},{\"question\":\"What is DRAGON and what does it achieve?\",\"answer\":\"DRAGON is a sentence-pair BERT multi-label classifier trained on large repository data and GitRanking topics, improving F1@5 by 11% over prior best work and maintaining performance even when many projects lack a README.\"}]","DOTTORATO DI RICERCA IN COMPUTER SCIENCE AND ENGINEERING - MACHINE LEARNING FOR SOFTWARE ENGINEERING - Thesis overview | PDF",1785811362,355,{"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},"phd-in-computer-science-and-engineering-machine-learning-for-software-engineering-thesis-overview","",{"@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/phd-in-computer-science-and-engineering-machine-learning-for-software-engineering-thesis-overview/122569/",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","2026-08-04",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 challenges motivate this ML4SE thesis?","Question",{"text":75,"@type":76},"The thesis targets three core issues: code is often reformatted/minified, benchmarks for repository recommendation are not standardized, and models must scale to archive-scale datasets such as Software Heritage.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the work represent code to improve robustness to transformations?",{"text":80,"@type":76},"It proposes a language-agnostic stylometric representation using Concrete Syntax Tree (CST) path-contexts, contrasted with traditional Abstract Syntax Tree (AST) approaches.",{"name":82,"@type":73,"acceptedAnswer":83},"What is DRAGON and what does it achieve?",{"text":84,"@type":76},"DRAGON is a sentence-pair BERT multi-label classifier trained on large repository data and GitRanking topics, improving F1@5 by 11% over prior best work and maintaining performance even when many projects lack a README.","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"]