[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-289794-105":59,"doc-detail-289794-en":131},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","an-empirical-study-on-data-distribution-aware-test-selection-for-deep-learning-enhancement","An Empirical Study on Data Distribution-Aware Test Selection for Deep Learning Enhancement","","Research focuses on continuous enhancement of deep neural networks under real-world distribution shifts, where labeling all new test data is labor-intensive. The work analyzes limitations of existing test selection metrics, including dependence on retraining setups, neglect of distribution shifts, and insufficient evaluation. It first conducts an empirical study of retraining process and data distribution effects, then introduces a distribution-aware test (DAT) selection metric that mitigates shift impact and delivers up to 5x improvements and 30.09% accuracy gains.",{"@graph":69,"@context":123},[70,84,106],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/an-empirical-study-on-data-distribution-aware-test-selection-for-deep-learning-enhancement/289794/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/an-empirical-study-on-data-distribution-aware-test-selection-for-deep-learning-enhancement/289794.png","ImageObject",300,407,{"name":92,"@type":93},"Mali","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-29","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",14,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What problem does the study address in deep learning testing?","Question",{"text":113,"@type":114},"It addresses how to continuously enhance deep neural networks when new test data follows distribution shifts, without manually labeling all collected test data.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"What are the key limitations of existing test selection metrics?",{"text":118,"@type":114},"Existing metrics are limited by using different retraining processes, ignoring data distribution shifts, and being insufficiently evaluated.",{"name":120,"@type":111,"acceptedAnswer":121},"What is the proposed solution and its reported benefit?",{"text":122,"@type":114},"The work proposes a distribution-aware test (DAT) selection metric. Retraining with both training and selected data outperforms using only the selected data, and DAT mitigates distribution-shift impact up to 5 times with 30.09% accuracy improvement in simulated and in-the-wild scenarios.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},289794,1789637123,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":140,"language":141,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":142,"faqs":143,"seo_title":144,"seo_description":67,"update_tm":130,"read_time":145,"preview_page_count":4},2336475104362,"https://ap-avatar.wpscdn.com/avatar/22000c4c46a41b752dd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1786595829695023868","An Empirical Study on Data Distribution-Aware Test Selection for Deep Learning Enhancement  \n1  \nQIANG HU, University of Luxembourg, Luxembourg YUEJUN GUO∗ , University of Luxembourg, Luxembourg MAXIME CORDY, University of Luxembourg, Luxembourg XIAOFEI XIE, Singapore Management University, Singapore LEI MA, University of Alberta, Canada  \nMIKE PAPADAKIS, University of Luxembourg, Luxembourg YVES LE TRAON, University of Luxembourg, Luxembourg  \nSimilar to traditional software that is constantly under evolution, deep neural networks (DNNs) need to evolve upon the rapid growth of test data for continuous enhancement, e.g., adapting to distribution shift in a new environment for deployment. However, it is labor-intensive to manually label all the collected test data. Test selection solves this problem by strategically choosing a small set to label. Via retraining with the selected set, DNNs will achieve competitive accuracy. Unfortunately, existing selection metrics involve three main limitations: 1) using different retraining processes; 2) ignoring data distribution shifts; 3) being insufficiently evaluated. To fill this gap, we first conduct a systemically empirical study to reveal the impact of the retraining process and data distribution on model enhancement. Then based on our findings, we propose a novel distribution-aware test (DAT) selection metric. Experimental results reveal that retraining using both the training and selected data outperforms using only the selected data. None of the selection metrics perform the best under various data distributions. By contrast, DAT effectively alleviates the impact of distribution shifts and outperforms the compared metrics by up to 5 times and 30.09% accuracy improvement for model enhancement on simulated and in-the-wild distribution shift scenarios, respectively.  \nCCS Concepts: • Software and its engineering → Empirical software validation; • Computing methodologies → Artificial intelligence.  \nAdditional Key Words and Phrases: deep learning testing, test selection, data distribution  \nACM Reference Format:  \nQiang Hu, Yuejun Guo∗, Maxime Cordy, Xiaofei Xie, Lei Ma, Mike Papadakis, and Yves Le Traon. 2022. An Empirical Study on Data Distribution-Aware Test Selection for Deep Learning Enhancement. ACM Trans. Softw. Eng. Methodol. 1, 1, Article 1 (January 2022), 31 pages. [https://doi.org/10.1145/3511598](https://doi.org/10.1145/3511598)  \n1 INTRODUCTION  \nDeep Neural Networks (DNNs) are increasingly integrated into large software systems in various applications, such as face recognition [39], autonomous vehicles [2], speech recognition [52], video gaming [44], and so on. Despite the impressive success and great potential of DNNs, there are crucial  \n∗ Corresponding Author.  \nAuthors’ addresses: Qiang Hu, University of Luxembourg, Luxembourg; Yuejun Guo∗, University of Luxembourg, Luxembourg; Maxime Cordy, University of Luxembourg, Luxembourg; Xiaofei Xie, Singapore Management University, Singapore; Lei Ma, University of Alberta, Canada; Mike Papadakis, University of Luxembourg, Luxembourg; Yves Le Traon, University of Luxembourg, Luxembourg.  \nPermission to make digital or hard copies of part or all ofthis work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s) .  \n© 2022 Copyright held by the owner/author(s) .  \n1049-331X/2022/1-ART1  \n[https://doi.org/10.1145/3511598](https://doi.org/10.1145/3511598)  \nACM Trans. Softw. Eng. Methodol., Vol. 1, No. 1, Article 1 . Publication date: January 2022 .  \n1:2 Qiang Hu, Yuejun Guo∗, Maxime Cordy, Xiaofei Xie, Lei Ma, Mike Papadakis, and Yves Le Traon  \naccidents caused by quality issues of deep learning (DL) systems, e.g., Tesla/Uber accidents [45] . Therefore, similar","cbCaihRsVZf5xJaN","https://ap.wps.com/l/cbCaihRsVZf5xJaN","pdf",1810528,31,"English","# Introduction\n## Motivation and problem context\n## Distribution shift challenges\n## Need for continuous DL enhancement","[{\"question\":\"What problem does the study address in deep learning testing?\",\"answer\":\"It addresses how to continuously enhance deep neural networks when new test data follows distribution shifts, without manually labeling all collected test data.\"},{\"question\":\"What are the key limitations of existing test selection metrics?\",\"answer\":\"Existing metrics are limited by using different retraining processes, ignoring data distribution shifts, and being insufficiently evaluated.\"},{\"question\":\"What is the proposed solution and its reported benefit?\",\"answer\":\"The work proposes a distribution-aware test (DAT) selection metric. Retraining with both training and selected data outperforms using only the selected data, and DAT mitigates distribution-shift impact up to 5 times with 30.09% accuracy improvement in simulated and in-the-wild scenarios.\"}]","An Empirical Study on Data Distribution-Aware Test Selection for Deep Learning Enhancement | PDF",78]