[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85680-en":3,"doc-seo-85680-105":29,"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":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":13,"seo_description":14,"update_tm":27,"read_time":28},85680,549758252649,"Ivy","https://ap-avatar.wpscdn.com/avatar/8000253669c5317157?_k=1778319167496531819",8,"Research & Report","FedCausal-Dyn: Causal-Dynamic Paradigm for Federated Learning under Dynamic Feature Drift","Dynamic feature drift in federated learning—where client and time-varying data distributions shift—causes ambiguous decision boundaries and degrades global model performance, and many methods assume static drift. FedCausal-Dyn introduces a causal-dynamic federated learning framework that separates domain-invariant causal features from spurious domain-specific variations using projection heads and adversarial training. Reliability-aware prototype aggregation and causal-feature guided collaborative regularization improve dynamic, stable global alignment. Experiments across three benchmarks show state-of-the-art accuracy and robustness, confirmed by ablations.","arXiv :2607 .09695v1 [ cs .LG] 20 Jun 2026  \nFedCausal-Dyn: A Causal-Dynamic Paradigm for Federated Learning under Dynamic Feature Drift  \nKaijie Chen Alex Johnson Maria Garcia Wei Zhang Daniel Kim  \nMindlab Mindlab Mindlab Mindlab Mindlab  \nAbstract  \nThis paper addresses the challenging problem of dynamic feature drift in federated learning, where data distributions evolve across clients and over time—a common scenario in real-world applications like financial technology. Existing approaches often assume static drift, limiting their effectiveness in non-stationary environments.  \nTo overcome this, we propose FedCausal-Dyn, a novel federated learning framework built on a causal-dynamic paradigm. Its key innovation is causal-domain feature separation, which disentangles domain-invariant causal features from spurious, domain-specific variations via specialized projection heads and adversarial training. This enables reliable and dynamic prototype aggregation, weighting local class prototypes by estimated reliability before global aggregation. We further introduce causal-feature guided collaborative regularization, unifying prototype contrastive alignment and domain invariance into a cohesive objective. Extensive experiments on three federated domain generalization benchmarks demonstrate that FedCausal-Dyn consistently achieves state-of-the-art performance, with the highest average accuracy and the most stable results. Ablation studies confirm each component’s critical contribution. Our work provides a robust and principled solution for federated learning under dynamic feature drift.  \n1 Introduction  \nIn computer vision, leveraging data from multiple sources is crucial for training robust models. However, direct data sharing raises significant privacy concerns. Federated Learning (FL) has emerged as a promising paradigm that enables collaborative model training while preserving data privacy by keeping raw data decentralized on local clients McMahan et al. [2017], Qu and Ma [2025] . Inspired by foundational federated learning approaches Wu et al. [2022], Qi et al. [2022], we build upon adaptive privacy-preserving mechanisms to develop robust solutions for heterogeneous data distributions. Despite its promise, FL systems face a critical challenge in data heterogeneity, where the data distributions across clients are statistically diverse (non-IID) . While much existing research focuses on label distribution skew, feature drift—a variation in the feature distributions for the same class across different clients due to disparate data collection environments—remains a prevalent yet underexplored problem Luo et al. [2025], Tian et al. [2025] . This drift leads to ambiguous decision boundaries and severely degrades global model performance.  \nCurrent strategies to mitigate feature drift primarily aim to align or decentralize the feature spaces across clients. These include employing client-specific normalization layers Li et al. [2021b], Lin [2025b] or sharing synthetic data and raw features for alignment Chen et al. [2023], Lin [2025a] . Building upon prior work in federated learning with differential privacy Wu et al. [2022], Lin [2025c] and intelligent access control frameworks Wang et al. [2023], He et al. [2025b], we extend these approaches to handle dynamic feature drift. However, such approaches often suffer from notable drawbacks: they may inadvertently suppress causally predictive features, lack robustness against unreliable client updates, and introduce potential privacy risks through the sharing of features or synthetic data Zhao et al. [2026], Dou et al. [2026a, 2025, 2026b], Zhao et al. [2026] .  \nPreprint.  \nTo overcome these limitations, we propose FedCausal-Dyn, a novel federated learning framework designed to address dynamic feature drift from a causal-dynamic perspective. Outperforming existing methods Wu et al. [2024c,a], Yang et al. [2025a], our framework explicitly models evolving data distributions and achieves su","cbCaiqgn5XxN5E72","https://ap.wps.com/l/cbCaiqgn5XxN5E72","pdf",5850030,1,18,"English","en",105,"# Introduction\n# Related Work\n# FedCausal-Dyn Framework\n# Experimental Results and Analysis\n# Ablation Studies\n# Limitations\n# Conclusion","[{\"question\":\"What problem does FedCausal-Dyn address in federated learning?\",\"answer\":\"It targets dynamic feature drift, where feature distributions for the same class evolve across clients and over time, leading to inconsistent representations and worse global performance.\"},{\"question\":\"How does the framework separate causal and spurious features?\",\"answer\":\"FedCausal-Dyn uses causal-domain feature separation with specialized projection heads and adversarial training to disentangle domain-invariant causal features from domain-specific spurious variations.\"},{\"question\":\"How are local prototypes combined during global aggregation?\",\"answer\":\"The method performs reliable and dynamic prototype aggregation by weighting local class prototypes according to estimated reliability before global aggregation.\"}]",1784205558,45,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"fedcausal-dyn-causal-dynamic-paradigm-for-federated-learning-under-dynamic-feature-drift","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/fedcausal-dyn-causal-dynamic-paradigm-for-federated-learning-under-dynamic-feature-drift/85680/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","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 FedCausal-Dyn address in federated learning?","Question",{"text":75,"@type":76},"It targets dynamic feature drift, where feature distributions for the same class evolve across clients and over time, leading to inconsistent representations and worse global performance.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the framework separate causal and spurious features?",{"text":80,"@type":76},"FedCausal-Dyn uses causal-domain feature separation with specialized projection heads and adversarial training to disentangle domain-invariant causal features from domain-specific spurious variations.",{"name":82,"@type":73,"acceptedAnswer":83},"How are local prototypes combined during global aggregation?",{"text":84,"@type":76},"The method performs reliable and dynamic prototype aggregation by weighting local class prototypes according to estimated reliability before global aggregation.","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":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":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]