[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85723-en":3,"doc-seo-85723-105":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},85723,4398048950312,"Violet","https://ap-avatar.wpscdn.com/avatar/400002538284de19e3c?_k=1778320343897328908",8,"Research & Report","JEPA for AI-Native 6G: Predictive Representations and Open Challenges","Sixth-generation (6G) networks are shifting to AI-native operation, embedding learning modules across the radio access network (RAN), edge, and core. This move demands learning from limited labels, heterogeneous wireless and network signals, partial observations, non-stationary propagation, and latency-constrained control loops. Joint-embedding predictive architecture (JEPA) enables self-supervised prediction of missing or future latent representations rather than raw reconstruction or contrastive negatives. A wireless-focused tutorial details JEPA training, tokenization and masking of CSI/beam/KPI/topology/sensing inputs, predictive encoder roles, and task-specific heads. A beam-management study shows improved label efficiency and robustness via an auxiliary future beam-energy target, followed by open challenges in multi-timescale prediction, action-conditioned modeling, distributed training, trustworthiness, deployment, benchmarking, and standardization.","JEPA for AI-Native 6G:  \nPredictive Representations and Open Challenges  \nSheikh Salman Hassan, Member, IEEE, Irshad A. Meer, Member, IEEE, Almoatssimbillah Saifaldawla, Member, IEEE, Yan Kyaw Tun, Senior Member, IEEE, Mustafa Ozger, Member, IEEE, Madyan Alsenwi, Member, IEEE, Nguyen Van Huynh, Member, IEEE, Woong-Hee Lee, Cedomir Stefanovic, Senior Member, IEEE, Mathini Sellathurai, Fellow, IEEE, Henk Wymeersch, Fellow, IEEE, and Tharmalingam Ratnarajah, Senior Member, IEEE  \narXiv :2607 .09798v 1 [ cs .LG] 9 Jul 2026  \nAbstract—Sixth-generation (6G) networks are moving toward AI-native operation, where learning modules are embedded across the radio access network (RAN), edge, and core. This transition requires learning from limited labels, heterogeneous wireless and network data, partial observations, non-stationary propagation, and latency-constrained control loops. Joint-embedding predictive architecture (JEPA) is a promising self-supervised paradigm for this setting because it predicts missing or future representations in latent space instead of reconstructing raw measurements or using contrastive negative samples. This article presents a wireless-oriented tutorial on JEPA for 6G intelligence. We define the JEPA training mechanism, describe how CSI, beam measurements, KPIs, topology graphs, and sensing observations can be tokenized and masked, and position the learned encoder asa predictive representation layer for RAN, O-RAN, edge, and core functions, with task-specific heads or controllers producing final decisions. Then we present an illustrative, beam-management case study suggesting that a wireless-aware target, specifically an auxiliary future beam-energy target during self-supervised pretraining, can improve label efficiency and robustness across shifted deployment conditions relative to a supervised source domain. Finally, we outline open challenges in multi-timescale prediction, action-conditioned modeling, distributed training, trustworthiness, efficient deployment, benchmarking, and standardization.  \nIndex Terms—6G, AI-native networks, self-supervised learning, joint-embedding predictive architecture (JEPA), predictive representations, beam management, O-RAN.  \nS. S. Hassan is with IDCoM, University of Edinburgh, Edinburgh EH9 3BF, UK. E-mail: [shassan@ed.ac.uk](shassan@ed.ac.uk).  \nI. A. Meer is with KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden. E-mail: [iameer@kth.se](iameer@kth.se).  \nA. Saifaldawla and M. Alsenwi are with the SnT, University of Luxembourg, L-1855 Luxembourg. E-mails: [moatssim.saifaldawla@uni.lu](moatssim.saifaldawla@uni.lu), [madyan.alsenwi@ieee.org](madyan.alsenwi@ieee.org).  \nY. K. Tun, M. Ozger, and C. Stefanovic are with the Department of Electronic Systems, Aalborg University, 9220 Aalborg, Denmark. E-mails: [ykt@es.aau.dk](ykt@es.aau.dk), [mozger@es.aau.dk](mozger@es.aau.dk), [cs@es.aau.dk](cs@es.aau.dk).  \nN. V. Huynh is with the School of Computer Science and Informatics, University of Liverpool, Liverpool L69 3DR, UK. E-mail: [huynh.nguyen@liverpool.ac.uk](huynh.nguyen@liverpool.ac.uk).  \nW.-H. Lee is with the Division of Electronics and Electrical Engineering, Dongguk University, Seoul 04620, Republic of Korea. Email: [woongheelee@dongguk.edu](woongheelee@dongguk.edu).  \nM. Sellathurai is with the School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, UK. E-mail:  \n[M.Sellathurai@hw.ac.uk](M.Sellathurai@hw.ac.uk).  \nH. Wymeersch is with the Department of Electrical Engineering, Chalmers University of Technology, 41296 Gteborg, Sweden. E-mail: [henkw@chalmers.se](henkw@chalmers.se).  \nT. Ratnarajah is with the Department of Electrical and Computer Engineering, San Diego State University, San Diego, CA 92182, USA. E-mail: [t.ratnarajah@ieee.org](t.ratnarajah@ieee.org).  \nI. Introduction & Background  \n6G networks are expected to move beyond improvements in throughput, latency, connectivity, and reliability toward artificial intel","cbCaiuZxWl93wIPI","https://ap.wps.com/l/cbCaiuZxWl93wIPI","pdf",620683,1,12,"English","en",105,"# Abstract\n## JEPA training and wireless tokenization\n## Beam-management case study\n## Open challenges","[{\"question\":\"Why is JEPA well-suited for AI-native 6G operation?\",\"answer\":\"JEPA supports self-supervised learning by predicting missing or future latent representations, fitting AI-native 6G scenarios with limited labels, heterogeneous observations, and latency-constrained control.\"},{\"question\":\"How are wireless and network observations used in JEPA training?\",\"answer\":\"The approach tokenizes and masks inputs such as CSI, beam measurements, KPIs, topology graphs, and sensing observations, enabling the learned encoder to act as a predictive representation layer for RAN, O-RAN, edge, and core tasks.\"},{\"question\":\"What is the purpose of the auxiliary future beam-energy target in the beam-management case study?\",\"answer\":\"It acts as a wireless-aware prediction target during self-supervised pretraining to improve label efficiency and robustness across shifted deployment conditions compared with a supervised source domain.\"}]",1784205812,30,{"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},"jepa-for-ai-native-6g-predictive-representations-and-open-challenges","",{"@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/jepa-for-ai-native-6g-predictive-representations-and-open-challenges/85723/",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},"Why is JEPA well-suited for AI-native 6G operation?","Question",{"text":75,"@type":76},"JEPA supports self-supervised learning by predicting missing or future latent representations, fitting AI-native 6G scenarios with limited labels, heterogeneous observations, and latency-constrained control.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How are wireless and network observations used in JEPA training?",{"text":80,"@type":76},"The approach tokenizes and masks inputs such as CSI, beam measurements, KPIs, topology graphs, and sensing observations, enabling the learned encoder to act as a predictive representation layer for RAN, O-RAN, edge, and core tasks.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the purpose of the auxiliary future beam-energy target in the beam-management case study?",{"text":84,"@type":76},"It acts as a wireless-aware prediction target during self-supervised pretraining to improve label efficiency and robustness across shifted deployment conditions compared with a supervised source domain.","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},[93,97,101,105,110,115,120,122,127,130,134],{"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":28,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]