[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85651-en":3,"doc-seo-85651-105":28,"detail-sidebar-cat-0-en-105":90},{"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":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":13,"seo_description":14,"update_tm":26,"read_time":27},85651,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","Separating Expert Retention from Autonomous Source Inference in Raw-ECG-Replay-Free Continual ECG Deployment","Multi-source ECG deployment must sometimes incorporate new data sources when earlier raw ECGs cannot be retained or replayed. Freezing a pretrained backbone and assigning isolated classifiers prevents parameter interference, yet real deployment still requires expert selection when source metadata are missing. The work studies this separation using IRFE-ECG, an incremental expert bank on frozen ECGFounder features with balanced-softmax experts and a retained-feature router, fused via a validation-calibrated top-2 margin rule.","Separating Expert Retention from Autonomous Source Inference in Raw-ECG-Replay-Free Continual ECG Deployment  \nYufan Lu 1 , Xinhui Liu 1 , Chenyang Xu 1 , Yuxi Zhou2,3,* , and Hao Wang 1,* ,  \n1Xidian University, China  \n2Department of Computer Science, Tianjin University of Technology, China  \n3DCST, BNRist, RIIT, Institute of Internet Industry, Tsinghua University, China  \nEmails: [24019100468@stu.xidian.edu.cn](24019100468@stu.xidian.edu.cn), [24049200220@stu.xidian.edu.cn](24049200220@stu.xidian.edu.cn), [xcy@ieee.org](xcy@ieee.org)  \njoy [yuxi@pku.edu.cn](yuxi@pku.edu.cn), [haow@ieee.org](haow@ieee.org)  \n*  \nCorresponding authors: Hao Wang and Yuxi Zhou  \narXiv :2607 .0 1674v 3 [ cs .AI] 13 Jul 2026  \nAbstract—In multi-source ECG deployment, models may need to incorporate new data sources when earlier raw ECGs cannot be retained or replayed. Freezing a pretrained backbone and assigning each source an isolated classifier prevents parameter interference, but deployment still requires selecting an expert when source metadata are unavailable. We study this distinction through IRFE-ECG, an incremental expert bank built on frozen 1024-dimensional ECGFounder features. Each arriving domain adds a balanced-softmax linear expert, while a lightweight router is fitted only on retained training features and domain labels from sources observed so far. A validation-calibrated margin rule fuses the two most likely experts instead of committing to a single routed expert.  \nOn CPSC, PTB-XL, Georgia, and Chapman-Shaoxing, sourceaware expert selection reaches 0.7915 ± 0.0036 Macro-F1 anda matched offline independent-head reference reaches 0.7885 ± 0.0009, showing that sequentially added isolated heads preserve source-aware performance relative to an offline independenthead control under the frozen-backbone design. Without source IDs, an MLP router reaches 0.7756 ± 0.0027 and top-2 margin fusion reaches 0.7782 ± 0.0022. The top-2 gain over hard MLP routing is small (+0 .0026), with a 95% confidence interval from paired bootstrap that includes zero. Across three domain orders, the top-2-to-oracle gap remains 0.0111–0.0133, supporting the interpretation that autonomous source inference, rather than expert-parameter retention, is the main remaining limitation within this protocol. All results should be interpreted as recordlevel benchmark evidence because reliable patient identifiers were unavailable. No raw ECGs are replayed, but frozen training features are retained for router updates; the method is therefore raw-ECG-replay-free but not memory-free. Code is available at [https://github.com/yufanlu221/IRFE-ECG](https://github.com/yufanlu221/IRFE-ECG).  \nIndex Terms—Electrocardiogram, continual learning, ECG foundation model, expert routing, domain-incremental learning, raw-ECG-replay-free learning, source inference.  \nI. INTRODUCTION  \nECG systems rarely remain confined to one stationary data distribution. Continual cardiac-signal studies have considered changes across diseases, time, modalities, and institutions [1], and multicenter ECG continual-learning work has highlighted data-governance and data-sharing constraints in sequential ECG adaptation [2] . A deployed ECG system may first receive data from one hospital or acquisition device and later  \nencounter records collected under different hardware, cohorts, or annotation protocols. Updating a shared model on each new source can overwrite earlier decision boundaries, while retaining or centralizing raw historical ECGs may be limited by data-governance and data-sharing constraints. Foundation models offer a useful alternative: a fixed representation can support lightweight domain-specific classifiers without repeatedly optimizing a large backbone.  \nParameter isolation alone, however, does not complete the deployment problem. If source metadata are known at inference time, selecting the corresponding frozen expert is straightforward, and earlier experts cannot be overwritten. If the source ","cbCaipheAbYNmm0u","https://ap.wps.com/l/cbCaipheAbYNmm0u","pdf",646326,1,"English","en",105,"# Introduction\n## Contributions\n# Related Work\n## ECG Foundation Models","[{\"question\":\"What problem does the paper distinguish in raw-ECG-replay-free continual ECG deployment?\",\"answer\":\"It separates (1) expert retention for each source from (2) autonomous source inference at test time when source metadata are unavailable.\"},{\"question\":\"How does IRFE-ECG update experts when a new domain arrives?\",\"answer\":\"Each new domain adds an isolated balanced-softmax linear expert, while the backbone remains frozen and earlier experts are never updated.\"},{\"question\":\"How is source inference performed when the source ID is unknown?\",\"answer\":\"A lightweight router is fitted using retained training features and domain labels observed so far, and performance is improved by a validation-calibrated top-2 margin fusion that combines the two most likely experts.\"}]",1784205367,20,{"code":4,"msg":29,"data":30},"ok",{"site_id":23,"language":22,"slug":31,"title":13,"keywords":32,"description":14,"schema_data":33,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":26},"separating-expert-retention-from-autonomous-source-inference-in-raw-ecg-replay-free-continual-ecg-deployment","",{"@graph":34,"@context":84},[35,52,67],{"@type":36,"itemListElement":37},"BreadcrumbList",[38,42,46,49],{"item":39,"name":40,"@type":41,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":43,"name":44,"@type":41,"position":45},"https://docshare.wps.com/document/","Document",2,{"item":47,"name":12,"@type":41,"position":48},"https://docshare.wps.com/document/research-report/",3,{"item":50,"name":13,"@type":41,"position":51},"https://docshare.wps.com/document/separating-expert-retention-from-autonomous-source-inference-in-raw-ecg-replay-free-continual-ecg-deployment/85651/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":61,"encodingFormat":59,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":39,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":20},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does the paper distinguish in raw-ECG-replay-free continual ECG deployment?","Question",{"text":74,"@type":75},"It separates (1) expert retention for each source from (2) autonomous source inference at test time when source metadata are unavailable.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does IRFE-ECG update experts when a new domain arrives?",{"text":79,"@type":75},"Each new domain adds an isolated balanced-softmax linear expert, while the backbone remains frozen and earlier experts are never updated.",{"name":81,"@type":72,"acceptedAnswer":82},"How is source inference performed when the source ID is unknown?",{"text":83,"@type":75},"A lightweight router is fitted using retained training features and domain labels observed so far, and performance is improved by a validation-calibrated top-2 margin fusion that combines the two most likely experts.","https://schema.org",{"og:url":50,"og:type":86,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":88,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":44,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":45,"doc_module":4,"doc_module_name":44,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":44,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":44,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":44,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":44,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":44,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":44,"category_name":124,"show_sort_weight":27,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":27,"doc_module":4,"doc_module_name":44,"category_name":127,"show_sort_weight":27,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":44,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":44,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]