[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85147-en":3,"doc-seo-85147-105":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":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":13,"seo_description":14,"update_tm":28,"read_time":29},85147,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","RUBRIC Realism-Utility Balanced Ranking for Imbalanced Classification","Class imbalance threatens accurate performance in risk-sensitive tasks like fraud detection and medical diagnosis, where minority samples are scarce but essential. Many oversampling pipelines rely on synthetic generation, yet they can add abundant low-quality candidates that warp decision boundaries, create artifacts, and harm generalization. RUBRIC introduces a generator-agnostic filtering method that ranks candidates via a realism–utility trade-off. Realism is measured by a learned discriminator, while utility reflects proximity to the decision boundary using a concave margin-based score. Theory and experiments show improved F1-macro and recall under fixed selection budgets while preserving similar ROC-AUC across generators, with explicit sensitivity analysis for λ and AUPRC behavior.","RUBRIC: Realism–Utility Balanced Ranking for Imbalanced  \nClassification  \nYanxuan Yu 1 , ∗ Dong Liu2 , ∗ Renata Borovica-Gajic3 Ying Nian Wu2  \n1 Columbia University 2 University of California, Los Angeles 3 University of Melbourne  \n[1](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ yy3523@columbia.edu](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ 2](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ pikeliu@ucla.edu](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ 3](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ renata.borovica@unimelb.edu.au](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ 2](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)[ ywu@stat.ucla.edu](1 yy3523@columbia.edu 2 pikeliu@ucla.edu 3 renata.borovica@unimelb.edu.au 2 ywu@stat.ucla.edu)  \n∗ Equal contribution.  \narXiv :2607 .098 16v 1 [ cs .LG] 10 Jul 2026  \nAbstract  \nClass imbalance poses a fundamental challenge in risk-sensitive applications such as fraud detection and medical diagnosis, where minority-class samples are scarce yet critical for accurate classification. Existing oversampling methods generate synthetic samples torebalance class distributions; however, they often produce large numbers of low-quality candidates that distort decision boundaries or introduce artifacts, leading to overfitting and degraded generalization. In this work, we introduce RUBRIC, a generator-agnostic filtering framework that formulates synthetic sample selection as a quality-over-quantity optimization problem. RUBRIC ranks candidates using a realism–utility trade-off: realism is quantified via a learned discriminator that distinguishes real from synthetic samples, while utility captures proximity to the decision boundary through a concave, marginbased scoring function. We show that, under mild regularity conditions, the proposed filtering strategy monotonically tightens the generalization bound for margin-based classifiers by jointly reducing distribution shift and suppressing near-negative tail contributions. Through extensive experiments on credit-card fraud detection and other imbalanced benchmarks, we demonstrate that RUBRIC improves F1-macro and recall while maintaining comparable ROC-AUC on several generators, with explicit λ-sensitivity analysis showing how users can recover AUPRC when ranking quality is prioritized.  \nCode and experimental scripts are publicly available at [https://github.com/zorinayu/RUBRIC](https://github.com/zorinayu/RUBRIC).  \n1 INTRODUCTION  \nSevere class imbalance is prevalent in risk screening and medical decision support He and Garcia (2009); Krawczyk (2016) . We focus on risk screening settings with public, highly imbalanced tabular benchmarks (e.g., fraud detection), where extreme skew is common and evaluation is reproducible. A standard response is to oversample the minority class, yet widely used generators can produce many low-quality candidates. For example, SMOTE Chawla et al. (2002) interpolates within local neighborhoods; without careful control, synthetic points may be unrealistic or push the decision boundary into regions unsupported by real data.  \nMost prior work addresses where to synthesize minority samples. Boundary-aware variants emphasize difficult regions, while GAN-based approaches seek greater sample realism through adversarial training Douzas and Bacao (2018); Fiore et al. (2019) . However, realism and downstream utility are not equivalent: a plausiblelooking sample may be uninformative or even harmful for learning, and aggressively boundary-focused synthesis can amplify distribution shift.  \nOur premise is orthogonal. Rather than designing new generators, we ask which synthetic sampl","cbCait0oh0J2Okhs","https://ap.wps.com/l/cbCait0oh0J2Okhs","pdf",8440186,3,1,40,"English","en",105,"# Introduction\n## Problem: class imbalance and oversampling risks\n## Prior work and its limitations\n## RUBRIC: generator-agnostic post-filtering\n## Contributions overview\n# Related Work\n## Minority generation methods","[{\"question\":\"What problem does RUBRIC address in imbalanced classification?\",\"answer\":\"RUBRIC targets class imbalance settings where oversampling with synthetic samples can introduce many low-quality candidates that distort decision boundaries and reduce generalization, especially in risk-sensitive tasks.\"},{\"question\":\"How does RUBRIC decide which synthetic samples to keep?\",\"answer\":\"RUBRIC ranks each candidate by two quantities: a realism score from a learned discriminator and a utility score tied to proximity to the decision boundary using a concave, margin-based scoring function, then selects a fixed-size subset to optimize the trade-off.\"},{\"question\":\"What benefits does RUBRIC show in experiments and metrics?\",\"answer\":\"Across credit-card fraud detection and other imbalanced benchmarks, RUBRIC improves recall and F1-macro under a fixed selection budget while keeping ROC-AUC comparable, and it provides λ-sensitivity analysis explaining how AUPRC changes when ranking emphasizes quality.\"}]",1784201379,101,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"rubric-realism-utility-balanced-ranking-for-imbalanced-classification","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"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":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/rubric-realism-utility-balanced-ranking-for-imbalanced-classification/85147/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does RUBRIC address in imbalanced classification?","Question",{"text":75,"@type":76},"RUBRIC targets class imbalance settings where oversampling with synthetic samples can introduce many low-quality candidates that distort decision boundaries and reduce generalization, especially in risk-sensitive tasks.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does RUBRIC decide which synthetic samples to keep?",{"text":80,"@type":76},"RUBRIC ranks each candidate by two quantities: a realism score from a learned discriminator and a utility score tied to proximity to the decision boundary using a concave, margin-based scoring function, then selects a fixed-size subset to optimize the trade-off.",{"name":82,"@type":73,"acceptedAnswer":83},"What benefits does RUBRIC show in experiments and metrics?",{"text":84,"@type":76},"Across credit-card fraud detection and other imbalanced benchmarks, RUBRIC improves recall and F1-macro under a fixed selection budget while keeping ROC-AUC comparable, and it provides λ-sensitivity analysis explaining how AUPRC changes when ranking emphasizes quality.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":21,"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":52,"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":22,"slug":118},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]