[{"data":1,"prerenderedAt":368},["ShallowReactive",2],{"publication-2025\u002Fdynamic-transfer-learning-switching-approach-using-resource-benchmark-in-edge-in-en":3,"publication-members":55},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"_hidden":6,"Intelligence authors":10,"authors_orcid":15,"year":20,"doi":21,"openalex_id":22,"venue":23,"abstract_screenshot":24,"keywords":25,"body":35,"_type":48,"_id":49,"_source":50,"_file":51,"_stem":52,"_extension":53,"locale":54},"\u002Fpublications\u002F2025\u002Fdynamic-transfer-learning-switching-approach-using-resource-benchmark-in-edge-in","2025",false,"","Dynamic Transfer Learning Switching Approach Using Resource Benchmark in Edge","Machine learning (ML) techniques are applied for profiling computing and processing resources data collected while running deep neural network models on edge devices. Adaptive deep neural network (DNN) model switching requires proper benchmarking for categorizing AI models based on their applications and computational resources enabled by their processing accelerators. Based on benchmark metrics, DNN models can be classified into tiny, low, small, medium, and large resources, then identify DNN models that perform well within resource constraints. Ensure efficient resource allocation, latency management, and trade-off between accuracy and resource. In this work, we propose a benchmark for edge transfer artificial intelligence learning service (TALS) that uses ML techniques. They aim at classifying DNN models by their target edge applications while running edge inferences. We used both unsupervised learning (UL) and supervised learning (SL) techniques to identify the most effective features for the TALS models and to benchmark the performance of edge devices. To achieve this, two approaches were investigated: first, determining features based on edge inference’s computing resources profiling using principal component analysis; and second, classifying the DNN models at the target application level using a regression approach based on historical resource utilization data. In addition, we propose a dynamic model transfer learning that switches between a set of pre-trained and optimized and quantized DNN models based on the cost function. ML techniques learn resource-aware prediction from new resource allocation data and ensure that the multicriteria switch cost selects the inference task models that meet the edge resource constraint requirements. The experimental results highlight a strong relationship between the supervised learning model and the clustering execution method. The dynamic switching approach on real edge devices demonstrates dynamic switching between models according to inference task complexity. We conclude that dynamic switching models allows to ensure smooth operation without overloading resources in edge intelligence.",[11,12,13,14],"Nkenyereye, Lionel","Rajkumar, Chellakannu","Lee, Boon Giin","Chung, Wan‐Young",[16,17,18,19],"0000-0001-6714-4402","0000-0002-0838-5350","0000-0001-5743-1010","0000-0002-0121-855X",2025,"https:\u002F\u002Fdoi.org\u002F10.1109\u002Fjiot.2025.3557408","W4409129983","IEEE Internet of Things Journal",null,[26,27,28,29,30,31,32,33,34],"Computer science","Benchmark (surveying)","Transfer of learning","Enhanced Data Rates for GSM Evolution","Artificial intelligence","Resource (disambiguation)","Distributed computing","Machine learning","Computer network",{"type":36,"children":37,"toc":45},"root",[38],{"type":39,"tag":40,"props":41,"children":42},"element","p",{},[43],{"type":44,"value":9},"text",{"title":7,"searchDepth":46,"depth":46,"links":47},2,[],"markdown","content:publications:2025:dynamic-transfer-learning-switching-approach-using-resource-benchmark-in-edge-in.md","content","publications\u002F2025\u002Fdynamic-transfer-learning-switching-approach-using-resource-benchmark-in-edge-in.md","publications\u002F2025\u002Fdynamic-transfer-learning-switching-approach-using-resource-benchmark-in-edge-in","md","en",[56,69,74,85,92,100,106,115,122,128,133,143,150,159,165,177,186,195,201,209,214,222,228,236,240,250,257,265,270,279,285,293,299,307,312,318,328,336,342,350,355,363],{"_path":57,"title":58,"name":59,"role":60,"email":24,"image":61,"category":62,"interests":63,"order":46,"_id":68},"\u002Fmembers\u002Fstaff\u002Falain-chong","Vice President for Global Affairs and Partnerships · Professor of Information Systems and Digital Innovation","Alain Chong","Deputy Director of Lab","assets\u002F8.png","staff",[64,65,66,67],"信息系统与运作管理","计算机科学与运筹学","Information Systems and Operations Management","Computer Science and Operations Research","content:members:staff:alain-chong.md",{"_path":57,"title":70,"role":71,"interests":72,"_id":73},"全球事务与合作副校长 · 信息系统与数字创新教授","实验室副主任",[64,65],"content:members:staff:alain-chong.zh-CN.md",{"_path":75,"title":76,"name":77,"role":78,"email":24,"image":79,"category":62,"interests":80,"order":83,"_id":84},"\u002Fmembers\u002Fstaff\u002Fanthony-belloti","Professor","Anthony Belloti","Core Member","assets\u002F41.png",[81,82],"Machine Learning and Credit Risk Model","Model Risks",9,"content:members:staff:anthony-belloti.md",{"_path":75,"title":86,"role":87,"interests":88,"_id":91},"计算机科学系教授","核心成员",[89,90],"机器学习与信用风险模型","模型风险","content:members:staff:anthony-belloti.zh-CN.md",{"_path":93,"title":94,"name":94,"role":78,"email":24,"image":95,"category":62,"interests":96,"order":98,"_id":99},"\u002Fmembers\u002Fstaff\u002Fboon-giin-lee","Boon Giin Lee","assets\u002F31.jpg",[97],"Intelligent Sensor and Extended Reality",11,"content:members:staff:boon-giin-lee.md",{"_path":93,"title":101,"role":87,"interests":102,"_id":105},"人机交互实验室负责人 · 计算机科学系副教授",[103,104],"人机交互 HCI","智能传感与扩展现实技术","content:members:staff:boon-giin-lee.zh-CN.md",{"_path":107,"title":108,"name":108,"role":109,"email":24,"image":110,"category":62,"interests":111,"order":113,"_id":114},"\u002Fmembers\u002Fstaff\u002Fcong-cao","Cong Cao","Direction Leader","assets\u002FCC.png",[112],"Science and technology policy and institutional reform",7,"content:members:staff:cong-cao.md",{"_path":107,"title":116,"name":117,"role":118,"interests":119,"_id":121},"宁波诺丁汉大学商学院创新学教授","曹聪","方向带头人",[120],"科技政策与体制改革","content:members:staff:cong-cao.zh-CN.md",{"_path":123,"title":124,"name":124,"role":78,"email":24,"image":125,"category":62,"order":126,"_id":127},"\u002Fmembers\u002Fstaff\u002Fdave-towey","Dave Towey","assets\u002F32.jpg",8,"content:members:staff:dave-towey.md",{"_path":123,"title":129,"role":87,"interests":130,"_id":132},"计算机科学系教授 · 计算机科学系主任",[131],"计算机科学与语言学","content:members:staff:dave-towey.zh-CN.md",{"_path":134,"title":135,"name":135,"role":78,"email":24,"image":136,"category":62,"interests":137,"order":141,"_id":142},"\u002Fmembers\u002Fstaff\u002Ffazl-ullah-khan","Fazl Ullah Khan","assets\u002F44.png",[138,139,140],"Computer Network","Computer Architecture and Network Security","Software Engineering",12,"content:members:staff:fazl-ullah-khan.md",{"_path":134,"title":144,"role":87,"interests":145,"_id":149},"计算机科学系助理教授 · IEEE 高级会员",[146,147,148],"计算机网络","计算机和网络安全","软件工程","content:members:staff:fazl-ullah-khan.zh-CN.md",{"_path":151,"title":152,"name":153,"role":78,"email":24,"image":154,"category":62,"interests":155,"order":157,"_id":158},"\u002Fmembers\u002Fstaff\u002Fheng-yu","Associate Professor","Heng Yu","assets\u002FHENGYU.png",[156],"Embedded Systems Design",17,"content:members:staff:heng-yu.md",{"_path":151,"title":160,"name":161,"role":87,"interests":162,"_id":164},"计算机科学系副教授","于恒",[163],"嵌入式系统设计","content:members:staff:heng-yu.zh-CN.md",{"_path":166,"title":152,"name":167,"role":78,"email":24,"image":168,"category":62,"interests":169,"order":175,"_id":176},"\u002Fmembers\u002Fstaff\u002Fheshan-du","Heshan Du","assets\u002Fhesahndu.png",[170,171,172,173,174],"Logic, Knowledge Representation and Reasoning","Geographic Information Systems","Operations Research","Machine Learning","Reinforcement Learning",20,"content:members:staff:heshan-du.md",{"_path":166,"title":160,"name":178,"role":87,"interests":179,"_id":185},"杜何珊",[180,181,182,183,184],"逻辑与知识表示","地理信息系统","运筹学","机器学习","强化学习","content:members:staff:heshan-du.zh-CN.md",{"_path":187,"title":188,"name":189,"role":78,"email":24,"image":190,"category":62,"interests":191,"order":193,"_id":194},"\u002Fmembers\u002Fstaff\u002Fhuan-jin","Assistant Professor","Huan Jin","assets\u002Fhuanjin.png",[192,173],"Optimisation",21,"content:members:staff:huan-jin.md",{"_path":187,"title":196,"name":197,"role":87,"interests":198,"_id":200},"计算机科学系助理教授","靳欢",[199,183],"优化","content:members:staff:huan-jin.zh-CN.md",{"_path":202,"title":152,"name":203,"role":109,"email":24,"image":204,"category":62,"interests":205,"order":207,"_id":208},"\u002Fmembers\u002Fstaff\u002Fjianfeng-ren","Jianfeng Ren","assets\u002F42.jpg",[173,206],"Computer Vision",3,"content:members:staff:jianfeng-ren.md",{"_path":202,"title":160,"name":210,"role":118,"interests":211,"_id":213},"任剑锋",[183,212],"计算机视觉","content:members:staff:jianfeng-ren.zh-CN.md",{"_path":215,"title":216,"name":216,"role":109,"email":24,"image":217,"category":62,"interests":218,"order":220,"_id":221},"\u002Fmembers\u002Fstaff\u002Fjiawei-li","Jiawei Li","assets\u002F11.png",[219],"Computer Science and Artificial Intelligence",15,"content:members:staff:jiawei-li.md",{"_path":215,"title":223,"name":224,"role":118,"interests":225,"_id":227},"计算机科学系助理教授 · 英国诺丁汉大学博士后","李家炜",[226],"计算机与人工智能","content:members:staff:jiawei-li.zh-CN.md",{"_path":229,"title":152,"name":230,"role":78,"email":24,"image":231,"category":62,"interests":232,"order":234,"_id":235},"\u002Fmembers\u002Fstaff\u002Fmatthew-pike","Matthew Pike","assets\u002F43.jpg",[233],"Digitalised Learning",16,"content:members:staff:matthew-pike.md",{"_path":229,"title":160,"role":87,"interests":237,"_id":239},[238],"数字化学习","content:members:staff:matthew-pike.zh-CN.md",{"_path":241,"title":188,"name":242,"role":78,"email":24,"image":243,"category":62,"interests":244,"order":248,"_id":249},"\u002Fmembers\u002Fstaff\u002Fning-xue","Ning Xue","\u002Fimages\u002Fuon-logo.png",[245,246,247],"Artificial Intelligence","Computational Intelligence","Combinatorial Optimization",13,"content:members:staff:ning-xue.md",{"_path":241,"title":196,"name":251,"role":87,"interests":252,"_id":256},"薛宁",[253,254,255],"人工智能","计算智能","组合优化","content:members:staff:ning-xue.zh-CN.md",{"_path":258,"title":188,"name":259,"role":78,"email":24,"image":260,"category":62,"interests":261,"order":263,"_id":264},"\u002Fmembers\u002Fstaff\u002Fqian-zhang","Qian Zhang","assets\u002Fqz.png",[262,206,173],"Image Processing",14,"content:members:staff:qian-zhang.md",{"_path":258,"title":196,"name":266,"role":87,"interests":267,"_id":269},"张茜",[268,212,183],"图像处理","content:members:staff:qian-zhang.zh-CN.md",{"_path":271,"title":76,"name":272,"role":273,"email":24,"image":274,"category":62,"interests":275,"orcid":276,"order":277,"_id":278},"\u002Fmembers\u002Fstaff\u002Fruibin-bai","Ruibin Bai","Director of Lab","assets\u002F38.png",[67],"0000-0003-1722-568X",1,"content:members:staff:ruibin-bai.md",{"_path":271,"title":280,"name":281,"role":282,"interests":283,"_id":284},"教授","白瑞斌","实验室主任",[65],"content:members:staff:ruibin-bai.zh-CN.md",{"_path":286,"title":287,"name":287,"role":109,"email":24,"image":288,"category":62,"interests":289,"order":291,"_id":292},"\u002Fmembers\u002Fstaff\u002Fsean-he","Sean He","assets\u002F39.png",[206,290,173],"Data Analytics",5,"content:members:staff:sean-he.md",{"_path":286,"title":294,"name":295,"role":118,"interests":296,"_id":298},"计算机科学系教授 · 国家级讲席学者","何祥健",[212,297,183],"数据分析","content:members:staff:sean-he.zh-CN.md",{"_path":300,"title":301,"name":301,"role":78,"email":24,"image":302,"category":62,"interests":303,"order":305,"_id":306},"\u002Fmembers\u002Fstaff\u002Ftianxiang-cui","Tianxiang Cui","assets\u002Ftianxiangcui.png",[246,304,173,174],"Operation Research",19,"content:members:staff:tianxiang-cui.md",{"_path":300,"title":144,"name":308,"role":87,"interests":309,"_id":311},"崔天翔",[254,310,183,184],"运筹研究","content:members:staff:tianxiang-cui.zh-CN.md",{"_path":313,"title":76,"name":314,"role":78,"email":24,"image":315,"category":62,"order":316,"_id":317},"\u002Fmembers\u002Fstaff\u002Fxiuping-hua","Xiuping Hua","assets\u002FxiupignHua.png",10,"content:members:staff:xiuping-hua.md",{"_path":313,"title":319,"name":320,"role":87,"interests":321,"_id":327},"金融、会计与经济系教授","华秀萍",[322,323,324,325,326,89],"资产定价","公司金融","衍生品","金融科技","创新金融和普惠金融","content:members:staff:xiuping-hua.zh-CN.md",{"_path":329,"title":76,"name":330,"role":78,"email":24,"image":331,"category":62,"interests":332,"order":334,"_id":335},"\u002Fmembers\u002Fstaff\u002Fying-weng","Ying Weng","assets\u002Fyingweng.png",[206,262,333],"IoT",4,"content:members:staff:ying-weng.md",{"_path":329,"title":86,"name":337,"role":87,"interests":338,"_id":341},"翁莹",[212,268,339,340],"物联网 IoT","无线网络安全与服务质量","content:members:staff:ying-weng.zh-CN.md",{"_path":343,"title":188,"name":344,"role":78,"email":24,"image":345,"category":62,"interests":346,"order":348,"_id":349},"\u002Fmembers\u002Fstaff\u002Fyuan-yao","Yuan Yao","assets\u002Fyuanyao.png",[347],"Autonomous Agents and Multi-Agent Systems",18,"content:members:staff:yuan-yao.md",{"_path":343,"title":196,"name":351,"role":87,"interests":352,"_id":354},"姚远",[353],"自主智能体与多智能体系统","content:members:staff:yuan-yao.zh-CN.md",{"_path":356,"title":188,"name":357,"role":109,"email":24,"image":358,"category":62,"interests":359,"order":361,"_id":362},"\u002Fmembers\u002Fstaff\u002Fzheng-lu","Zheng Lu","assets\u002F13.png",[360],"Computer Science",6,"content:members:staff:zheng-lu.md",{"_path":356,"title":196,"name":364,"role":118,"interests":365,"_id":367},"卢正",[366],"计算机科学","content:members:staff:zheng-lu.zh-CN.md",1782639480429]