[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117752-en":3,"doc-seo-117752-105":30,"detail-sidebar-cat-0-en-105":83},{"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},117752,1099513958607,"Jiven","https://ap-avatar.wpscdn.com/avatar/100002390cf8733938c?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778829742770036399",8,"Research & Report","Data Models for Dataset Drift Controls in Machine Learning With Images","Camera images are ubiquitous in machine learning research and critical for services spanning medicine and environmental surveying, yet model deployment is often limited by robustness concerns. Performance drops commonly arise from discrepancies between training and deployment data, known as dataset drift. Existing robustness validation methods typically do not rely on explicit models of the underlying data, which limits physically faithful drift test cases and precise specifications of data conditions to avoid. This work couples robustness validation with physical optics using differentiable data models to enable drift synthesis, drift forensics, and drift adjustment, including public datasets and released code.","Data Models for Dataset Drift Controls in Machine Learning With Images  \nLuis Oala∗ [luis.oala@hhi.frauhofer.de](luis.oala@hhi.frauhofer.de)  \nFraunhofer HHI  \nMarco Aversa∗ [marco.aversa@dotphoton.com](marco.aversa@dotphoton.com)  \nDotphoton AG and University of Glasgow  \nKurt Willis [kurt.willis@hhi.frauhofer.de](kurt.willis@hhi.frauhofer.de)  \nFraunhofer HHI  \nGabriel Nobis [gabriel.nobis@hhi.frauhofer.de](gabriel.nobis@hhi.frauhofer.de)  \nFraunhofer HHI  \nYoan Neuenschwander yoan. neuenschwander@hesge. ch  \nHEPIA/HES-SO  \nMichèle Buck [michele.kyncl@tum.de](michele.kyncl@tum.de)  \nKlinikum rechts der Isar  \nChristian Matek [christian.matek@helmholtz-muenchen.de](christian.matek@helmholtz-muenchen.de)  \nHelmholtz Zentrum Munich  \nJérôme Extermann [jerome.extermann@hesge.ch](jerome.extermann@hesge.ch)  \n[HEPIA/HES-SO](HEPIA/HES-SO)  \nEnrico Pomarico [enrico.pomarico@hesge.ch](enrico.pomarico@hesge.ch)  \n[HEPIA/HES-SO](HEPIA/HES-SO)  \nWojciech Samek [wojciech.samek@hhi.frauhofer.de](wojciech.samek@hhi.frauhofer.de)  \nFraunhofer HHI  \nRoderick Murray-Smith [roderick.murray-smith@glasgow.ac.uk](roderick.murray-smith@glasgow.ac.uk)  \nUniversity of Glasgow  \nChristoph Clausen [christoph.clausen@dotphoton.com](christoph.clausen@dotphoton.com)  \nDotphoton AG  \nBruno Sanguinetti [bruno.sanguinetti@dotphoton.com](bruno.sanguinetti@dotphoton.com)  \nDotphoton AG  \n􀀃 Equal contribution  \nAbstract  \nCamera images are ubiquitous in machine learning research. They also play a central role in the delivery of important services spanning medicine and environmental surveying. However, the application of machine learning models in these domains has been limited because of robustness concerns. A primary failure mode are performance drops due to di􀀛erences between the training and deployment data. While there are methods to prospectively validate the robustness of machine learning models to such dataset drifts, existing approaches do not account for explicit models of the primary object of interest: the data. This makes it di􀀞cult to create physically faithful drift test cases or to provide precise speci􀀜cations of data models that should be avoided during the deployment of a machine learning model. In this study, we demonstrate how these shortcomings can be overcome by pairing machine learning robustness validation with physical optics. We examine the role raw sensor data and di􀀛erentiable data models can play in controlling performance risks related to image dataset drift. The 􀀜ndings are distilled into three applications. First, drift synthesis enables the controlled generation of physically faithful drift test cases. The results for absolute and relative changes in task model performance obtained with our method diverge markedly from an augmentation testing alternative that is not physically faithful. Second, the gradient connection between machine learning model and our data models allows for drift forensics that can be used to specify performance-sensitive data models which should be avoided during deployment of a machine learning model. Third, drift adjustment opens up the possibility for processing adjustmentsin the face of drift. This can lead to speed up and stabilization of classi􀀜er training at a margin of up to 20% in validation accuracy. Alongside our data model code we release two datasets to the public that we collected as part of this work. In total, the two datasets, Raw-Microscopy and Raw-Drone, comprise 1,488 scienti􀀜cally calibrated reference raw sensor measurements, 8,928 raw intensity variations as well as 17,856 images processed through our data models with twelve di􀀛erent con􀀜gurations. A guide to access the open code and datasets is available at [https://github.com/aiaudit-org/raw2logit](https://github.com/aiaudit-org/raw2logit).  \n1 Introduction  \nIn this study we demonstrate how explicit data models for images can be constructed to enjoy advanced controls in the validation of machine learning model robustness to datase","cbCaiu3Z7raSPNJ7","https://ap.wps.com/l/cbCaiu3Z7raSPNJ7","pdf",34986715,1,56,"English","en",105,"# Abstract\n# 1 Introduction\n## Motivation and failure modes\n## Dataset drift and robustness validation","[{\"question\":\"What datasets and code are released, and why?\",\"answer\":\"The authors release two public datasets (Raw-Microscopy and Raw-Drone) plus access to code through an open repository, collected and processed as part of the study to support the proposed protocols.\"}]","Data Models for Dataset Drift Controls in Machine Learning With Images | PDF",1785679367,141,{"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":78,"head_meta":80,"extra_data":82,"updated_unix":28},"data-models-for-dataset-drift-controls-in-machine-learning-with-images","",{"@graph":36,"@context":77},[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/data-models-for-dataset-drift-controls-in-machine-learning-with-images/117752/",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],{"name":72,"@type":73,"acceptedAnswer":74},"What datasets and code are released, and why?","Question",{"text":75,"@type":76},"The authors release two public datasets (Raw-Microscopy and Raw-Drone) plus access to code through an open repository, collected and processed as part of the study to support the proposed protocols.","Answer","https://schema.org",{"og:url":52,"og:type":79,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":81,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":84},[85,89,93,97,102,107,112,115,120,123,127],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":86,"show_sort_weight":87,"slug":88},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":90,"show_sort_weight":91,"slug":92},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Exam",70,"exam",{"id":98,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},5,"Comic",60,"comic",{"id":103,"doc_module":4,"doc_module_name":46,"category_name":104,"show_sort_weight":105,"slug":106},6,"Technology",50,"technology",{"id":108,"doc_module":4,"doc_module_name":46,"category_name":109,"show_sort_weight":110,"slug":111},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":113,"slug":114},30,"research-report",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},9,"Religion & Spirituality",20,"religion-spirituality",{"id":118,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":118,"slug":122},"World Cup","world-cup",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":124,"slug":126},10,"Lifestyle","lifestyle",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":98,"slug":130},19,"General","general"]