[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-126879-en":3,"doc-seo-126879-105":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},126879,1099523885336,"Violet","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","The Evolution of Pro-Kremlin Propaganda From a Machine Learning and Linguistics Perspective - Abstract and Related Work","In the Russo-Ukrainian war, propaganda is produced by Russian state-run news outlets for both international and domestic audiences. Its content and form evolve during the conflict, creating difficulty for machine learning–based content moderation systems when training data diverges from current news. This follow-up study re-evaluates BERT and SVM classifiers trained on early-2022 news articles and Telegram posts using new 2023 data. It analyzes model errors and compares narrative shifts to explain performance drops.","Secondary Publication  \nSolopova, Veronika; Benzmüller, Christoph; Landgraf, Tim  \nThe Evolution of Pro-Kremlin Propaganda From a Machine Learning and Linguistics Perspective  \nDate of secondary publication: 07.09.2023  \nVersion of Record (Published Version), Conferenceobject Persistent identifier: urn:nbn:de:bvb:473-irb-904758  \nPrimary publication  \nSolopova, Veronika; Benzmüller, Christoph; Landgraf, Tim (2023): „The Evolution of  \nPro-Kremlin Propaganda From a Machine Learning and Linguistics Perspective“. In: Mariana Romanyshyn (Hrsg. ), Proceedings of the Second Ukrainian Natural Language Processing Workshop (UNLP 2023), Dubrovnik, S. 40-48.  \nLegal Notice  \nThis work is protected by copyright and/or the indication of a licence. You are free to use this work in any way permitted by the copyright and/or the licence that applies to your usage. For other uses, you must obtain permission from the rights-holder(s) .  \nThis document is made available under a Creative Commons license.  \nThe license information is available online:  \n[https://creativecommons.org/l](https://creativecommons.org/l)icenses/by/4 .0/legalcode  \nThe Evolution of Pro-Kremlin Propaganda From a Machine Learning and  \nLinguistics Perspective  \nVeronika Solopova  \nFreie Universität Berlin, Germany [veronika.solopova@fu-berlin.de](veronika.solopova@fu-berlin.de)  \nChristoph Benzmüller  \nUniversität Bamberg, Germany Freie Universität Berlin, Germany  \n[christoph.benzmueller@uni-bamberg.de](christoph.benzmueller@uni-bamberg.de)  \nTim Landgraf  \nFreie Universität Berlin, Germany [tim.landgraf@fu-berlin.de](tim.landgraf@fu-berlin.de)  \nAbstract  \nIn the Russo-Ukrainian war, propaganda is produced by Russian state-run news outlets for both international and domestic audiences. Its content and form evolve and change with time as the war continues. This constitutes a challenge to content moderation tools based on machine learning when the data used for training and the current news start to differ significantly. In this follow-up study, we evaluate our previous BERT and SVM models that classify Pro-Kremlin propaganda from a Pro-Western stance, trained on the data from news articles and telegram posts at the start of 2022, on the new 2023 subset. We examine both classifiers’errors and perform a comparative analysis of these subsets to investigate which changes in narratives provoke drops in performance.  \n1 Introduction and Related Work  \nFake news has been shown to evolve over time (Adriani, 2019) . A piece of news is often modified as it spreads online by malicious users who twist the original information (Guo et al., 2021), while an imperfect replication process by other users leads to further distortion (Zellers et al., 2019) . Guo et al.(2021) showed that the disinformation techniques, parts of speech, and keywords stayed consistent during the evolution process, while the text similarity and sentiment changed. Moreover, according to their scoring, the distance between the fake and evolved fake news was more prominent than between the truth and the initial fake news. The evolved ones sound more objective and cheerful and are more difficult to detect. Jang et al., 2018 also observed significant differences between real and fake news regarding evolution patterns. They found that fake news tweets underwent a more sig  \nnificant number of modifications over the spreading process.  \nInn case of fake news and disinformation originating in state-run outlets, we talk about propaganda. In this and previous studies, we focus on  \nRussian propaganda. (Kendall, 2014 ; Chee, 2017 ; Parlapiano and Lee, 2018) . It has been shown that the Russian Presidential Administration exercises coordinated control over media advertising budgets and editorial content whilst maintaining an illusion of media freedom by letting a small number of minor independent media outlets operate (Lange-Ionatamišvili, 2015) . Hence, the adaptations to Kremlin’s political agenda are an additional factor t","cbCaidUMZwzFIn0p","https://ap.wps.com/l/cbCaidUMZwzFIn0p","pdf",1246316,1,10,"English","en",105,"# Abstract\n# Introduction and Related Work\n## Evolution of fake news and disinformation\n## Russian state-run propaganda and narrative shifts\n## Prior work on pro-Kremlin propaganda identification","[{\"question\":\"Why is classifying pro-Kremlin propaganda challenging over time?\",\"answer\":\"As the war continues, propaganda content and form evolve, causing a mismatch between early training data and later news. This divergence reduces the effectiveness of machine learning classifiers for moderation tasks.\"},{\"question\":\"Which models are evaluated in the follow-up study?\",\"answer\":\"The study evaluates two classifiers: multilingual BERT and a Support Vector Machine (SVM) trained with linguistic features and manipulative terms.\"},{\"question\":\"What data split does the study compare to understand performance changes?\",\"answer\":\"It compares models trained on early 2022 news articles and Telegram posts against a new 2023 subset, then examines narrative changes that lead to drops in performance.\"}]","The Evolution of Pro-Kremlin Propaganda From a Machine Learning and Linguistics Perspective - Abstract and Related Work | PDF",1785935380,25,{"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},"the-evolution-of-pro-kremlin-propaganda-from-a-machine-learning-and-linguistics-perspective-abstract-and-related-work","",{"@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/the-evolution-of-pro-kremlin-propaganda-from-a-machine-learning-and-linguistics-perspective-abstract-and-related-work/126879/",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","2026-08-05",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},"Why is classifying pro-Kremlin propaganda challenging over time?","Question",{"text":75,"@type":76},"As the war continues, propaganda content and form evolve, causing a mismatch between early training data and later news. This divergence reduces the effectiveness of machine learning classifiers for moderation tasks.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which models are evaluated in the follow-up study?",{"text":80,"@type":76},"The study evaluates two classifiers: multilingual BERT and a Support Vector Machine (SVM) trained with linguistic features and manipulative terms.",{"name":82,"@type":73,"acceptedAnswer":83},"What data split does the study compare to understand performance changes?",{"text":84,"@type":76},"It compares models trained on early 2022 news articles and Telegram posts against a new 2023 subset, then examines narrative changes that lead to drops in performance.","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},[93,97,101,105,110,115,120,123,128,131,134],{"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":21,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]