In the network slicing environment, SDN and NFV technologies improve the utilization of network resources. However, these technologies lead to the greater dynamic nature of network resources and bring greater challenges to network fault management. Firstly, based on the analysis of network elements in network slicing environment, the fault management platform architecture in network slicing environment is proposed. Secondly, in order to improve the efficiency of the fault management mechanism, the knowledge graph is applied to the fault management field and the fault management knowledge graph is constructed. The functions that the knowledge framework can realize include fault prediction, elastic strategy, fault location, and fault self-healing. Finally, according to the requirements of fault management, a fault management question answering system based on knowledge graph is designed. From the two dimensions of business feasibility and implementation feasibility, the feasibility of the fault management platform is verified.
{"title":"Fault Management Platform based on Knowledge Graph in Network Slicing Environment","authors":"Detai Pan, Yunzhou Dong, Minglu Qi, Qiujia Fu, Zhengdong Lin, Guanliang Chen, Yingkun Liao, Peng Lin","doi":"10.1109/BMSB58369.2023.10211138","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211138","url":null,"abstract":"In the network slicing environment, SDN and NFV technologies improve the utilization of network resources. However, these technologies lead to the greater dynamic nature of network resources and bring greater challenges to network fault management. Firstly, based on the analysis of network elements in network slicing environment, the fault management platform architecture in network slicing environment is proposed. Secondly, in order to improve the efficiency of the fault management mechanism, the knowledge graph is applied to the fault management field and the fault management knowledge graph is constructed. The functions that the knowledge framework can realize include fault prediction, elastic strategy, fault location, and fault self-healing. Finally, according to the requirements of fault management, a fault management question answering system based on knowledge graph is designed. From the two dimensions of business feasibility and implementation feasibility, the feasibility of the fault management platform is verified.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"637 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78987306","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211244
Ran An, Zexuan Jing, Quan Zhou, Junsheng Mu
With the continuous combination of the localization field and AI methods, the accuracy of localization services has been improving. For example, in the field of indoor Localization based on WiFi fingerprint signals can be used for indoor Localization, monitoring and tracking tasks, but still faces many unsolved problems, such as poor Localization accuracy, vague floor Localization, high consumption of algorithm training samples, and data security risks. In this paper, Dual-indicator Localization Network designed based on Multitask Learning is considered for indoor Dual-indicator real-time localization based on WiFi fingerprint signals. Simulation experiments are also designed, and the analysis of the results from several dimensions such as confusion matrix, t-SNE graph, and model scoring criterion shows that the proposed DLnet network is much better than the traditional Machine Learning methods with a balance of localization accuracy and localization complexity.
{"title":"Indoor Dual-indicator Precision Localization Network based on Multitask Learning","authors":"Ran An, Zexuan Jing, Quan Zhou, Junsheng Mu","doi":"10.1109/BMSB58369.2023.10211244","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211244","url":null,"abstract":"With the continuous combination of the localization field and AI methods, the accuracy of localization services has been improving. For example, in the field of indoor Localization based on WiFi fingerprint signals can be used for indoor Localization, monitoring and tracking tasks, but still faces many unsolved problems, such as poor Localization accuracy, vague floor Localization, high consumption of algorithm training samples, and data security risks. In this paper, Dual-indicator Localization Network designed based on Multitask Learning is considered for indoor Dual-indicator real-time localization based on WiFi fingerprint signals. Simulation experiments are also designed, and the analysis of the results from several dimensions such as confusion matrix, t-SNE graph, and model scoring criterion shows that the proposed DLnet network is much better than the traditional Machine Learning methods with a balance of localization accuracy and localization complexity.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"17 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78974815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211097
Huai Yang, Xingyu Hou, Xia Jing, Junsheng Mu
This paper provides an overview of broadcast service technology enabled by integrated communication and navigation. The communication-navigation integration technique is based on orthogonal time frequency space (OTFS) modulation technique, OTFS is a novel two-dimensional modulation technique that can be used in high Doppler spread scenarios. OTFS has novel and important features designed in the delay-Doppler domain. In the case of high Doppler spread, OTFS can provide more significant performance improvements than OFDM. In addition, the broadcast technology empowered by the integrated communication and navigation technology is then introduced. In this paper, the existing research results are reviewed and summarized to derive the advantages of communication-navigation integrated technology-empowered broadcasting service techniques.
{"title":"Broadcast Service Technology Empowered by Integrated Communication and Navigation","authors":"Huai Yang, Xingyu Hou, Xia Jing, Junsheng Mu","doi":"10.1109/BMSB58369.2023.10211097","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211097","url":null,"abstract":"This paper provides an overview of broadcast service technology enabled by integrated communication and navigation. The communication-navigation integration technique is based on orthogonal time frequency space (OTFS) modulation technique, OTFS is a novel two-dimensional modulation technique that can be used in high Doppler spread scenarios. OTFS has novel and important features designed in the delay-Doppler domain. In the case of high Doppler spread, OTFS can provide more significant performance improvements than OFDM. In addition, the broadcast technology empowered by the integrated communication and navigation technology is then introduced. In this paper, the existing research results are reviewed and summarized to derive the advantages of communication-navigation integrated technology-empowered broadcasting service techniques.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"54 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78871840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211140
Zhongmiao Kang, Zanhong Wu, Peiming Zhang, Donghai Huang, Chang Liu
The power communication network uses SDN and NFV technology to improve the utilization of network resources, while bringing the characteristics of on-demand expansion and dynamic migration to the network. These new features pose new challenges to network fault management. To solve the problem that the existing research results lack an efficient fault management platform and mechanism for front-line maintenance personnel, this paper proposes a micro-service-based fault management platform and mechanism. Firstly, according to the technical composition and characteristics of the power communication network, the power communication network architecture is designed. Secondly, in order to make the network fault management mechanism programmable, fast adaptive, and dynamic, a fault management platform architecture based on microservices is designed according to the power communication network model, and the operation mechanisms of four business services, data collection, data processing, data analysis, and data view, are designed in detail. Finally, the feasibility of the platform and mechanism is analyzed from the four levels of the technical architecture.
{"title":"Microservice-based Fault Management Platform and Mechanism in Power Communication Network","authors":"Zhongmiao Kang, Zanhong Wu, Peiming Zhang, Donghai Huang, Chang Liu","doi":"10.1109/BMSB58369.2023.10211140","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211140","url":null,"abstract":"The power communication network uses SDN and NFV technology to improve the utilization of network resources, while bringing the characteristics of on-demand expansion and dynamic migration to the network. These new features pose new challenges to network fault management. To solve the problem that the existing research results lack an efficient fault management platform and mechanism for front-line maintenance personnel, this paper proposes a micro-service-based fault management platform and mechanism. Firstly, according to the technical composition and characteristics of the power communication network, the power communication network architecture is designed. Secondly, in order to make the network fault management mechanism programmable, fast adaptive, and dynamic, a fault management platform architecture based on microservices is designed according to the power communication network model, and the operation mechanisms of four business services, data collection, data processing, data analysis, and data view, are designed in detail. Finally, the feasibility of the platform and mechanism is analyzed from the four levels of the technical architecture.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"35 6 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82799733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211632
Fangning Shi, Zhong Na, Zhifeng Gao
With the integration of new generation information technology and industry, the value of industrial data has been highly valued. Meanwhile, enterprises are also facing various data security threats and urgently need to establish a trusted and controllable industrial data circulation platform and ecosystem, to eliminate enterprise concerns, unleash the vitality of data element sharing and circulation. This paper uses principal component analysis to classify, filter, transform or process the dataset, supporting the configuration of warning rules. Subsequently, based on a configurable rule engine, security policy verification and abnormal behavior warning are executed to achieve real-time supervision and tracking of the entire data circulation process, timely discover and warn of risk behaviors, avoid the risk of data leakage and protect data circulation security.
{"title":"Configurable Rule Engine for Industrial Data Flow Scenarios","authors":"Fangning Shi, Zhong Na, Zhifeng Gao","doi":"10.1109/BMSB58369.2023.10211632","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211632","url":null,"abstract":"With the integration of new generation information technology and industry, the value of industrial data has been highly valued. Meanwhile, enterprises are also facing various data security threats and urgently need to establish a trusted and controllable industrial data circulation platform and ecosystem, to eliminate enterprise concerns, unleash the vitality of data element sharing and circulation. This paper uses principal component analysis to classify, filter, transform or process the dataset, supporting the configuration of warning rules. Subsequently, based on a configurable rule engine, security policy verification and abnormal behavior warning are executed to achieve real-time supervision and tracking of the entire data circulation process, timely discover and warn of risk behaviors, avoid the risk of data leakage and protect data circulation security.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"51 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80993065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211224
Shiwei Li, Fangqing Tan, Qiang Liu
Combining cell-free massive multiple-input multiple-output (CF-mMIMO) and mobile edge computing (MEC) facilitates the processing of compute-intensive and latency-sensitive tasks in the distributed IoT. For MEC-enabled CF-mMIMO system, this paper designs a task offloading strategy for local computing and multi-access points (APs) collaboration. Under energy constraints, we aim to minimize the latency of computing offloading. According to the different data size of each user and the service of APs, the graph neural network method is adopted to deal with the link prediction between the user and APs.
{"title":"Mobile Edge Computing Tasking Offloading Strategy in Cell-Free Massive MIMO with Graph Neural Network","authors":"Shiwei Li, Fangqing Tan, Qiang Liu","doi":"10.1109/BMSB58369.2023.10211224","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211224","url":null,"abstract":"Combining cell-free massive multiple-input multiple-output (CF-mMIMO) and mobile edge computing (MEC) facilitates the processing of compute-intensive and latency-sensitive tasks in the distributed IoT. For MEC-enabled CF-mMIMO system, this paper designs a task offloading strategy for local computing and multi-access points (APs) collaboration. Under energy constraints, we aim to minimize the latency of computing offloading. According to the different data size of each user and the service of APs, the graph neural network method is adopted to deal with the link prediction between the user and APs.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"194 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81069758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Recently, the popularity of Internet of Things (IoT) devices has brought massive amounts of sensing data to edge networks. How to use distributed sensing data to train artificial intelligence (AI) models for ubiquitous intelligent wireless sensing in IoT, while considering privacy is an open problem. In this paper, we propose an edge intelligence (EI) and blockchain powered device-free wireless sensing framework to supply IoT applications in edge networks. We design a cross-domain wireless sensing scheme for human-computer interaction by adopting adversarial transfer learning in this framework and verify the effectiveness of the method.
{"title":"Blockchain-Aided Distributed Device-Free Wireless Sensing with IoT Devices in Edge Network","authors":"Yanxi Xie, Ziyue Li, Chaoyi Li, Yonghui Zhu, Hao Zhang","doi":"10.1109/BMSB58369.2023.10211278","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211278","url":null,"abstract":"Recently, the popularity of Internet of Things (IoT) devices has brought massive amounts of sensing data to edge networks. How to use distributed sensing data to train artificial intelligence (AI) models for ubiquitous intelligent wireless sensing in IoT, while considering privacy is an open problem. In this paper, we propose an edge intelligence (EI) and blockchain powered device-free wireless sensing framework to supply IoT applications in edge networks. We design a cross-domain wireless sensing scheme for human-computer interaction by adopting adversarial transfer learning in this framework and verify the effectiveness of the method.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"31 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81342134","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211170
Yunqi Li, J. Liu, Bin Jiang, Chan-Ming Yang, Qingtian Wang
With an increasing number of Mobile Users (MUs), Multi-access edge computing (MEC) has become a bottleneck in resource limitation. Serverless edge computing (SEC) is a promising approach to effectively alleviate the shortage of MEC. However, existing research on SEC focus on the operating mode of the SEC server, they ignore the interaction between SEC and MU. To this end, we propose a Stackelberg game approach to maximize the utility of each MU. We present the model of the Stackelberg game and propose an iterative algorithm as the solution. We also consider the impact of the function resource pool and using renewable energy on SEC. In particular, when a function that required by a MU is not stored in this SEC, it downloads the function from could with extra cost. Meanwhile, the SEC has a lower cost by using harvested energy rather than purchasing from the grid. Simulation results show that the proposed scheme is efficient in terms of SEC’s profit and MU’s demand. Moreover, both MUs and SECs gain benefits from renewable energy.
{"title":"Cost Minimization in Serverless Computing with Energy Harvesting SECs","authors":"Yunqi Li, J. Liu, Bin Jiang, Chan-Ming Yang, Qingtian Wang","doi":"10.1109/BMSB58369.2023.10211170","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211170","url":null,"abstract":"With an increasing number of Mobile Users (MUs), Multi-access edge computing (MEC) has become a bottleneck in resource limitation. Serverless edge computing (SEC) is a promising approach to effectively alleviate the shortage of MEC. However, existing research on SEC focus on the operating mode of the SEC server, they ignore the interaction between SEC and MU. To this end, we propose a Stackelberg game approach to maximize the utility of each MU. We present the model of the Stackelberg game and propose an iterative algorithm as the solution. We also consider the impact of the function resource pool and using renewable energy on SEC. In particular, when a function that required by a MU is not stored in this SEC, it downloads the function from could with extra cost. Meanwhile, the SEC has a lower cost by using harvested energy rather than purchasing from the grid. Simulation results show that the proposed scheme is efficient in terms of SEC’s profit and MU’s demand. Moreover, both MUs and SECs gain benefits from renewable energy.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"20 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81425423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211288
S. Jeon, Seongman Min, Dawoon Chung, Kangsoo Kim, Jahoon Ku, Sunhyung Kwon, Sung-Ik Park
Long-term Time Series Data from ATSC 3.0 single frequency broadcast network operation is important to understand anomaly patterns by measurement metrics because it provides a comprehensive view of the performance and behavior of the ATSC 3.0 network over time. This information can help identify and analyze patterns, trends, and outliers in the data, which can provide valuable insights into the health and stability of the network. By understanding these anomaly patterns, engineers and technicians can improve the network’s performance, troubleshoot issues, and make informed decisions about future upgrades and maintenance.
{"title":"Simple Anomaly Detection Technique from Long-term Time-series Data by ATSC 3.0 Single Frequency Broadcast Network Monitoring System","authors":"S. Jeon, Seongman Min, Dawoon Chung, Kangsoo Kim, Jahoon Ku, Sunhyung Kwon, Sung-Ik Park","doi":"10.1109/BMSB58369.2023.10211288","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211288","url":null,"abstract":"Long-term Time Series Data from ATSC 3.0 single frequency broadcast network operation is important to understand anomaly patterns by measurement metrics because it provides a comprehensive view of the performance and behavior of the ATSC 3.0 network over time. This information can help identify and analyze patterns, trends, and outliers in the data, which can provide valuable insights into the health and stability of the network. By understanding these anomaly patterns, engineers and technicians can improve the network’s performance, troubleshoot issues, and make informed decisions about future upgrades and maintenance.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"14 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87652502","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-06-14DOI: 10.1109/BMSB58369.2023.10211135
Biao Zhang, Siya Xu, Xusong Qiu, Jingyue Tian
Federated learning (FL) is an emerging distributed machine learning paradigm that emphasizes user privacy. The majority of current federated learning systems are oriented for single task while participants in the Internet of Vehicles (IoV) demand to train different models for multiple intelligent services simultaneously. As one of transfer learning methods, multi-task learning (MTL) has the potential to integrate with federated learning to realize personalized local training. However, the existing federated multi-task learning (FMTL) algorithms are faced with the problems of high implementation complexity and communication overhead. To solve the above issues, we propose a logical-cluster-based personalized federated multi-task learning framework named pFMTL. In the framework, the multi-task model is decomposed into a basic module for extracting features and K task-specific modules for outputting inferences. We leverage logical clusters and multi-task learning to enhance the personalization and generalization capability of task models, respectively. To improve the communication efficiency further, we also design a module-wise task scheduling strategy, which supports both user module scheduling and cluster aggregation scheduling to ensure the convergence of multi-task model with less communication overhead. Finally, the simulation results imply that pFMTL can increase task accuracy and reduce communication latency compared with other benchmarks.
{"title":"Logical-Cluster-Based Personalized Federated Multi-Task Learning for Internet of Vehicles","authors":"Biao Zhang, Siya Xu, Xusong Qiu, Jingyue Tian","doi":"10.1109/BMSB58369.2023.10211135","DOIUrl":"https://doi.org/10.1109/BMSB58369.2023.10211135","url":null,"abstract":"Federated learning (FL) is an emerging distributed machine learning paradigm that emphasizes user privacy. The majority of current federated learning systems are oriented for single task while participants in the Internet of Vehicles (IoV) demand to train different models for multiple intelligent services simultaneously. As one of transfer learning methods, multi-task learning (MTL) has the potential to integrate with federated learning to realize personalized local training. However, the existing federated multi-task learning (FMTL) algorithms are faced with the problems of high implementation complexity and communication overhead. To solve the above issues, we propose a logical-cluster-based personalized federated multi-task learning framework named pFMTL. In the framework, the multi-task model is decomposed into a basic module for extracting features and K task-specific modules for outputting inferences. We leverage logical clusters and multi-task learning to enhance the personalization and generalization capability of task models, respectively. To improve the communication efficiency further, we also design a module-wise task scheduling strategy, which supports both user module scheduling and cluster aggregation scheduling to ensure the convergence of multi-task model with less communication overhead. Finally, the simulation results imply that pFMTL can increase task accuracy and reduce communication latency compared with other benchmarks.","PeriodicalId":13080,"journal":{"name":"IEEE international Symposium on Broadband Multimedia Systems and Broadcasting","volume":"1 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2023-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88805899","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：