计算机科学 ›› 2022, Vol. 49 ›› Issue (11): 266-276.doi: 10.11896/jsjkx.211000067
胡朝霞1, 胡海周1, 蒋从锋1, 万健2
HU Zhao-xia1, HU Hai-zhou1, JIANG Cong-feng1and WAN Jian2
摘要: 边缘智能指利用人工智能算法为网络边缘设备提供数据分析能力的一种服务形式。然而,边缘计算环境比云计算更加复杂和多变。在构建边缘智能的过程中存在很多问题,例如缺乏量化的评价标准、异构计算平台、复杂的网络拓扑、不断变化的用户需求等,其中比较突出的是算法模型的高资源需求与边缘设备资源储备低之间的矛盾。机器学习是边缘智能的主要工作负载,它需要大量的计算资源,然而边缘设备的计算资源有限,两者的供求关系并不匹配,边缘智能负载的部署和优化成为了一个难题。因此,针对边缘智能负载性能优化问题,文中提出了基于负载特征的边缘智能性能优化CECI(Cloud -Edge Collaborative Inference)策略,从模型选择、批量自适应调整和云边协同方面对不同机器学习负载进行了优化。在模型选择方面,使用基于目标权重的模型自适应选择策略,实现在多个条件约束下,综合权衡多个性能优化目标的效果。在批量自适应调整方面,提出了基于开销反馈的批量自适应调整算法,使得模型在运行时能够达到更好的性能。在云边协同方面,通过结合网络状态和用户时延要求设计出了云边协同策略,进而达到了动态利用云端计算资源的效果。实验结果表明,与云智能相比,所提出的基于负载特征的边缘智能能够缩短50.79%的程序运行时间,降低了42.46%的系统能耗,并提升了4.52%的模型准确率。
中图分类号:
[1]LU Y,CHEN Y,LI T,et al.Embedded FPGA convolutionalneuralnetwork construction method for edge computing[J].Journal of Computer Research and Development,2018,55(3):551-562. [2]DWORK C,MCSHERRY F,NISSIM K,et al.Calibrating noise to sensitivity in private data analysis[C]//Theory of cryptography conference.2006:265-284. [3]FREDRIKSON M,JHA S,RISTENPART T.Model inversionattacks that exploit confidence information and basic countermeasures[C]// Proceedings of the 22nd ACM SIGSAC Confe-rence on Computer and Communications Security.2015:1322-1333. [4]DU M,WANG K,XIA Z,et al.Differential privacy preserving of training model in wireless big data with edge computing[J].IEEE transactions on big data,2018,6(2):283-295. [5]LIU T T,YANG C Y,SUO S Q,et al.Edge Intelligence inWireless Communication[J].Signal Processing,2020,36(11):1789-1803. [6]REN J,GAO L,YU J L,et al.Energy-efficient deep learning task scheduling strategy for edge devices[J].Chinese Journal of Computers,2020,43(3):440-452. [7]BHATTACHARYA S,LANE N D.Sparsification and separation of deep learning layers for constrained resource inference on wearables[C]//Proceedings of the 14th ACM Conference on Embedded Network Sensor Systems CD-ROM.2016:176-189. [8]VALERIO L,PASSARELLA A,CONTI M.A communication efficient distributed learning framework for smart environments [J].Pervasive and Mobile Computing,2017,41:46-68. [9]HSIEH K,HARLAP A,VIJAYKUMAR N,et al.Gaia:Geo-distributed machine learning approaching speeds[C]//14th Symposium on Networked Systems Design and Implementation(NSD).2017:629-647. [10]OGDEN S S,GUO T.MODI:Mobile deep inference made efficient by edge computing[C]// Workshop on Hot Topics in Edge Computing(HotEdge 18).2018:30-36. [11]HU C,BAO W,WANG D,et al.Dynamic adaptive DNN surgeryfor inference acceleration on the edge[C]//IEEE INFOCOM 2019-IEEE Conference on Computer Communications.2019:1423-1431. [12]ZHAO Z,BARIJOUGH K M,GERSTLAUER A.DeepThings:Distributed adaptive deep learning inference on resource-constrained IoT edge clusters[J].IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems,2018,37(11):2348-2359. [13]SHENG T,FENG C,ZHUO S,et al.A quantization-friendlyseparable convolution for mobilenets[C]//2018 1st Workshop on Energy Efficient Machine Learning and Cognitive Computing for Embedded Applications(EMC2).2018:14-18. [14]SANTOS A G,DE SOUZA C O,ZANCHETTIN C,et al.Reducing squeezenet storage size with depthwise separable convolutions[C]//2018 International Joint Conference on Neural Networks(IJCNN).2018:1-6. [15]CERUTTI G,PRASAD R,BRUTTI A,et al.Neural network distillation on IoT platforms for sound event detection[C]//Interspeech.2019:3609-3613. [16]YAO S,ZHAO Y,ZHANG A,et al.DeepIoT:Compressing deep neural network structures for sensing systems with a compressor-critic framework[C]//Proceedings of the 15th ACM Confe-rence on Embedded Network Sensor Systems.2017:1-14. [17]HAN S,KANG J,MAO H,et al.Ese:Efficient speech recognition engine with sparse lstm on fpga[C]//Proceedings of the 2017 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays.2017:75-84. [18]LIU S,LIN Y,ZHOU Z, et al.On-demand deep model compression for mobile devices:A usage-driven model selection framework[C]//Proceedings of the 16th Annual International Conference on Mobile Systems,Applications,and Services.2018:389-400. [19]HARDY C,LE MERRER E,SERICOLA B.Distributed deep learning on edge-devices:Feasibility via adaptive compression[C]//2017 IEEE 16th International Symposium on Network Computing and Applications(NCA).2017:1-8. [20]DEAN J,CORRADO G,MONGA R,et al.Large scale distributed deep networks[C]//Advances in Neural Information Processing Systems.2012:1223-1231. [21]YU Z,HU J,MIN G,et al.Federated learning based proactive content caching in edge computing[C]// 2018 IEEE Global Communications Conference(GLOBECOM).2018:1-6. [22]WANG S,TUOR T,SALONIDIS T,et al.When edge meets learning:Adaptive control for resource-constrained distributed machine learning[C]// IEEE INFOCOM 2018-IEEE Confe-rence on Computer Communications.2018:63-71. [23]FANG B,ZENG X,ZHANG M.Nestdnn:Resource-aware multi-tenant on-device deep learning for continuous mobile vision[C]//Proceedings of the 24th Annual International Conference on Mobile Computing and Networking.2018:115-127. [24]TAYLOR B,MARCO V S,WOLFF W,et al.Adaptive deeplearning model selection on embedded systems[J].ACM SIGPLAN Notices,2018,53(6):31-43. [25]KANG Y,HAUSWALD J,GAO C,et al.Neurosurgeon:Colla-borative intelligence between the cloud and mobile edge[J].ACM SIGARCH Computer Architecture News,2017,45(1):615-629. [26]MAO J,CHEN X,NIXON K W,et al.MODNN:Local distributed mobile computing system for deep neural network[C]// Design,Automation & Test in Europe Conference & Exhibition(DATE).2017:1396-1401. [27]ImageNet website update[OL].2021-06-28.https://www.image-net.org/. |
[1] | 刘兴光, 周力, 刘琰, 张晓瀛, 谭翔, 魏急波. 基于边缘智能的频谱地图构建与分发方法 Construction and Distribution Method of REM Based on Edge Intelligence 计算机科学, 2022, 49(9): 236-241. https://doi.org/10.11896/jsjkx.220400148 |
[2] | 孙慧婷, 范艳芳, 马孟晓, 陈若愚, 蔡英. VEC中基于动态定价的车辆协同计算卸载方案 Dynamic Pricing-based Vehicle Collaborative Computation Offloading Scheme in VEC 计算机科学, 2022, 49(9): 242-248. https://doi.org/10.11896/jsjkx.210700166 |
[3] | 于滨, 李学华, 潘春雨, 李娜. 基于深度强化学习的边云协同资源分配算法 Edge-Cloud Collaborative Resource Allocation Algorithm Based on Deep Reinforcement Learning 计算机科学, 2022, 49(7): 248-253. https://doi.org/10.11896/jsjkx.210400219 |
[4] | 李梦菲, 毛莺池, 屠子健, 王瑄, 徐淑芳. 基于深度确定性策略梯度的服务器可靠性任务卸载策略 Server-reliability Task Offloading Strategy Based on Deep Deterministic Policy Gradient 计算机科学, 2022, 49(7): 271-279. https://doi.org/10.11896/jsjkx.210600040 |
[5] | 袁昊男, 王瑞锦, 郑博文, 吴邦彦. 基于Fabric的电子病历跨链可信共享系统设计与实现 Design and Implementation of Cross-chain Trusted EMR Sharing System Based on Fabric 计算机科学, 2022, 49(6A): 490-495. https://doi.org/10.11896/jsjkx.210500063 |
[6] | 方韬, 杨旸, 陈佳馨. D2D辅助移动边缘计算下的卸载策略优化 Optimization of Offloading Decisions in D2D-assisted MEC Networks 计算机科学, 2022, 49(6A): 601-605. https://doi.org/10.11896/jsjkx.210200114 |
[7] | 刘漳辉, 郑鸿强, 张建山, 陈哲毅. 多无人机使能移动边缘计算系统中的计算卸载与部署优化 Computation Offloading and Deployment Optimization in Multi-UAV-Enabled Mobile Edge Computing Systems 计算机科学, 2022, 49(6A): 619-627. https://doi.org/10.11896/jsjkx.210600165 |
[8] | 谢万城, 李斌, 代玥玥. 空中智能反射面辅助边缘计算中基于PPO的任务卸载方案 PPO Based Task Offloading Scheme in Aerial Reconfigurable Intelligent Surface-assisted Edge Computing 计算机科学, 2022, 49(6): 3-11. https://doi.org/10.11896/jsjkx.220100249 |
[9] | 周天清, 岳亚莉. 超密集物联网络中多任务多步计算卸载算法研究 Multi-Task and Multi-Step Computation Offloading in Ultra-dense IoT Networks 计算机科学, 2022, 49(6): 12-18. https://doi.org/10.11896/jsjkx.211200147 |
[10] | 彭冬阳, 王睿, 胡谷雨, 祖家琛, 王田丰. 视频缓存策略中QoE和能量效率的公平联合优化 Fair Joint Optimization of QoE and Energy Efficiency in Caching Strategy for Videos 计算机科学, 2022, 49(4): 312-320. https://doi.org/10.11896/jsjkx.210800027 |
[11] | 张海波, 张益峰, 刘开健. 基于NOMA-MEC的车联网任务卸载、迁移与缓存策略 Task Offloading,Migration and Caching Strategy in Internet of Vehicles Based on NOMA-MEC 计算机科学, 2022, 49(2): 304-311. https://doi.org/10.11896/jsjkx.210100157 |
[12] | 林潮伟, 林兵, 陈星. 边缘环境下基于模糊理论的科学工作流调度研究 Study on Scientific Workflow Scheduling Based on Fuzzy Theory Under Edge Environment 计算机科学, 2022, 49(2): 312-320. https://doi.org/10.11896/jsjkx.201000102 |
[13] | 袁昕旺, 谢智东, 谭信. 无人机边缘计算中的资源管理优化研究综述 Survey of Resource Management Optimization of UAV Edge Computing 计算机科学, 2022, 49(11): 234-241. https://doi.org/10.11896/jsjkx.211100015 |
[14] | 李晓波, 陈鹏, 帅彬, 夏云霓, 李建岐. 边缘环境下轨迹预测性感知的在线边缘服务分配 Novel Predictive Approach to Trajectory-aware Online Edge Service Allocation in Edge Environment 计算机科学, 2022, 49(11): 277-283. https://doi.org/10.11896/jsjkx.211100029 |
[15] | 梁俊斌, 张海涵, 蒋婵, 王天舒. 移动边缘计算中基于深度强化学习的任务卸载研究进展 Research Progress of Task Offloading Based on Deep Reinforcement Learning in Mobile Edge Computing 计算机科学, 2021, 48(7): 316-323. https://doi.org/10.11896/jsjkx.200800095 |
|