[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121708-en":3,"doc-seo-121708-105":30,"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":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},121708,962075006959,"Anda","https://ap-avatar.wpscdn.com/avatar/e0002397efbe92a78e?_k=1776741047341049297",8,"Research & Report","NETWORK-AUTHORIZED SPLIT DATA RATE IN ARTIFICIAL INTELLIGENCE/MACHINE LEARNING RENDERING - Paper","Convolutional Neural Network (CNN) inference can be accelerated on mobile devices by splitting the model between the device and a network server. The device keeps privacy- and delay-sensitive parts while offloading computation- and energy-intensive parts. Methods describe how a network authorizes which CNN split level to use for AI/ML rendering, based on policy, real-time radio conditions, and client identification, while enforcing QoS-subscribed limits to meet bandwidth and latency constraints.","Technical Disclosure Commons  \nDefensive Publications Series  \nMay 2023  \nNETWORK-AUTHORIZED SPLIT DATA RATE IN ARTIFICIAL INTELLIGENCE/MACHINE LEARNING RENDERING Sri Gundavelli  \nVimal Srivastava  \nFollow this and additional works at: [https://www.tdcommons.org/dpubs_series](https://www.tdcommons.org/dpubs_series)  \nRecommended Citation  \nGundavelli, Sri and Srivastava, Vimal, \"NETWORK-AUTHORIZED SPLIT DATA RATE IN ARTIFICIAL INTELLIGENCE/MACHINE LEARNING RENDERING\", Technical Disclosure Commons,(May 10, 2023)  \n[https://www.tdcommons.org/dpubs_series/5881](https://www.tdcommons.org/dpubs_series/5881)  \nThis work is licensed under a Creative Commons Attribution 4.0 License.  \nThis Article is brought to you for free and open access by Technical Disclosure Commons. It has been accepted for inclusion in Defensive Publications Series by an authorized administrator of Technical Disclosure Commons.  \nNETWORK-AUTHORIZED SPLIT DATA RATE IN ARTIFICIAL  \nINTELLIGENCE/MACHINE LEARNING RENDERING  \nAUTHORS:  \nSri Gundavelli  \nVimal Srivastava  \nABSTRACT  \nConvolutional Neural Network (CNN) models are widely used for image/video recognition tasks on mobile devices. The CNN is split into two parts-the computationintensive and energy-intensive parts are offloaded to a network server while the privacysensitive and delay-sensitive parts are maintained at the end device. Techniques described herein allow a network to control the split for a user device in artificial intelligence (AI)/machine learning (ML) rendering based on a number offactors.  \nDETAILED DESCRIPTION  \nImage and video data are the largest sources of data on today’s Internet. Videos account for over 70% of daily Internet traffic. CNN models have been widely used for image/video recognition tasks (e.g., image classification, image segmentation, object localization and detection, face authentication, action recognition, enhanced photography, virtual reality (VR)/augmented reality (AR), video games, etc.) on mobile devices. Many references have shown that AI/ML inference for image processing with device-network synergy can alleviate the pressure of computation, memory footprint, storage, power and required data rate on devices, reduce end-to-end latency and energy consumption, and improve the end-to-end accuracy, efficiency and privacy when compared to the local execution approach on either side.  \nAccording to the current image recognition task and environment, the CNN is split into two parts. The intention is to offload the computation-intensive and energy-intensive parts to a network server while maintaining the privacy-sensitive and delay-sensitive parts at the end device. The device executes the inference up to a specific CNN layer and sends the intermediate data to the network server. The network server runs through the remaining CNN layers.  \n1 6837  \nPublished by Technical Disclosure Commons, 2023 2  \nConsider a case in which a device wants to split at layer 1 out of a total of 17 layers of CNN AlexNet processing. While this reduces the processing on the device (which requires less battery power), it increases the bandwidth requirement to transmit the data to the application server for the rest of the processing. Splitting is a function of policy, network load conditions, and user equipment (UE) wireless conditions, so it is logical that the network will decide where to do the split. Since the splitting is dependent on the network, the split is a 5G core (5GC) property and the 5GC controls the splitting. Techniques described herein discuss methods for allowing a network to decide which split model a device should use.  \nThere are many aspects to consider when deciding the split level. Figure 1, below, illustrates a message flow in which a decision of where to do the split is made based on a  \nnumber offactors.  \nFigure 1: Diagram of Example Message Flow for Determining Split  \nAs illustrated in Figure 1, the network should consider a maximum amount of bandwidth the network c","cbCaijtXnxG01r1a","https://ap.wps.com/l/cbCaijtXnxG01r1a","pdf",174249,1,4,"English","en",105,"# Abstract\n# Detailed Description\n## CNN split for mobile AI/ML inference\n## Device-layer splitting decision model\n## Factors: bandwidth, radio conditions, client identification\n## Network authorization based on QoS","[{\"question\":\"What does the proposed method do with CNN models in mobile AI/ML rendering?\",\"answer\":\"It splits the CNN into two parts: offloads computation- and energy-intensive layers to a network server while keeping privacy- and delay-sensitive layers on the end device.\"},{\"question\":\"Why is bandwidth a key issue when splitting at an early CNN layer?\",\"answer\":\"Splitting earlier reduces device processing power but increases the bandwidth needed to transmit intermediate data to the application server for the remaining layers.\"},{\"question\":\"How does the network decide and authorize the split level?\",\"answer\":\"The network considers policy bandwidth limits, real-time radio conditions, and client identification of the split point, then authorizes the split according to the UE-subscribed QoS values and related factors.\"}]","NETWORK-AUTHORIZED SPLIT DATA RATE IN ARTIFICIAL INTELLIGENCE/MACHINE LEARNING RENDERING - Paper | PDF",1785806401,10,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":28},"network-authorized-split-data-rate-in-artificial-intelligencemachine-learning-rendering-paper","",{"@graph":36,"@context":84},[37,53,67],{"@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":21},"https://docshare.wps.com/document/network-authorized-split-data-rate-in-artificial-intelligencemachine-learning-rendering-paper/121708/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What does the proposed method do with CNN models in mobile AI/ML rendering?","Question",{"text":74,"@type":75},"It splits the CNN into two parts: offloads computation- and energy-intensive layers to a network server while keeping privacy- and delay-sensitive layers on the end device.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"Why is bandwidth a key issue when splitting at an early CNN layer?",{"text":79,"@type":75},"Splitting earlier reduces device processing power but increases the bandwidth needed to transmit intermediate data to the application server for the remaining layers.",{"name":81,"@type":72,"acceptedAnswer":82},"How does the network decide and authorize the split level?",{"text":83,"@type":75},"The network considers policy bandwidth limits, real-time radio conditions, and client identification of the split point, then authorizes the split according to the UE-subscribed QoS values and related factors.","https://schema.org",{"og:url":52,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,127,130,133],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":29,"doc_module":4,"doc_module_name":46,"category_name":131,"show_sort_weight":29,"slug":132},"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":46,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]