Cost-Driven Off-Loading for DNN-Based Applications Over Cloud, Edge, and End Devices

Lin, Bing; Huang, Yinhao; Zhang, Jianshan; Hu, Junqin; Chen, Xing; Li, Jun

doi:10.1109/tii.2019.2961237

Cited by 137 publications

(38 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Huang et al [41] proposed a DeePar framework which can exploit all the available resources from the device, the edge, and the cloud to improve the overall inference performance. Lin et al [42] proposed a cost-driven offloading strategy based on a self-adaptive particle swarm optimization (PSO) algorithm using the genetic algorithm (GA) operators (PSO-GA) to minimize the system cost during offloading DNN layers over the cloud, edge, and devices.…”

Section: Hierarchy-basedmentioning

confidence: 99%

HierTrain: Fast Hierarchical Edge AI Learning With Hybrid Parallelism in Mobile-Edge-Cloud Computing

Liu

Chen

Zhou

et al. 2020

IEEE Open J. Commun. Soc.

View full text Add to dashboard Cite

Nowadays, deep neural networks (DNNs) are the core enablers for many emerging edge AI applications. Conventional approaches for training DNNs are generally implemented at central servers or cloud centers for centralized learning, which is typically time-consuming and resource-demanding due to the transmission of a large number of data samples from the edge device to the remote cloud. To overcome these disadvantages, we consider accelerating the learning process of DNNs on the Mobile-Edge-Cloud Computing (MECC) paradigm. In this paper, we propose HierTrain, a hierarchical edge AI learning framework, which efficiently deploys the DNN training task over the hierarchical MECC architecture. We develop a novel hybrid parallelism method, which is the key to HierTrain, to adaptively assign the DNN model layers and the data samples across the three levels of the edge device, edge server and cloud center. We then formulate the problem of scheduling the DNN training tasks at both layer-granularity and sample-granularity. Solving this optimization problem enables us to achieve the minimum training time. We further implement a hardware prototype consisting of an edge device, an edge server and a cloud server, and conduct extensive experiments on it. Experimental results demonstrate that HierTrain can achieve up to 6.9× speedups compared to the cloud-based hierarchical training approach. INDEX TERMS Edge AI, deep learning, fast model training, mobile-edge-cloud computing.

show abstract

Section: Hierarchy-basedmentioning

confidence: 99%

HierTrain: Fast Hierarchical Edge AI Learning With Hybrid Parallelism in Mobile-Edge-Cloud Computing

Liu

Chen

Zhou

et al. 2020

IEEE Open J. Commun. Soc.

View full text Add to dashboard Cite

show abstract

“…Moreover, MEC intensifies the processing capacities of mobile networks by deploying computational and storage resources at the network edge. Therefore, it would be an efficient way to alleviate the traffic load of core networks by partitioning DNNs and offloading DNN layers over the cloud, edge, and IoT devices [6].…”

Section: Introductionmentioning

confidence: 99%

Energy-Efficient Offloading for DNN-Based Smart IoT Systems in Cloud-Edge Environments

Chen

Zhang

Lin

et al. 2022

IEEE Trans. Parallel Distrib. Syst.

Self Cite

145

View full text Add to dashboard Cite

Deep Neural Networks (DNNs) have become an essential and important supporting technology for smart Internet-of-Things (IoT) systems. Due to the high computational costs of large-scale DNNs, it might be infeasible to directly deploy them in energy-constrained IoT devices. Through offloading computation-intensive tasks to the cloud or edges, the computation offloading technology offers a feasible solution to execute DNNs. However, energy-efficient offloading for DNN based smart IoT systems with deadline constraints in the cloud-edge environments is still an open challenge. To address this challenge, we first design a new system energy consumption model, which takes into account the runtime, switching, and computing energy consumption of all participating servers (from both the cloud and edge) and IoT devices. Next, a novel energy-efficient offloading strategy based on a Self-adaptive Particle Swarm Optimization algorithm using the Genetic Algorithm operators (SPSO-GA) is proposed. This new strategy can efficiently make offloading decisions for DNN layers with layer partition operations, which can lessen the encoding dimension and improve the execution time of SPSO-GA. Simulation results demonstrate that the proposed strategy can significantly reduce energy consumption compared to other classic methods.

show abstract

“…As a central focus of innovation and entrepreneurship education, students have the most critical say in their satisfaction with the teaching quality of the Fundamentals of Entrepreneurship curriculum (Harkema and Schout, 2010;Huanget al, 2019;Lin et al, 2020). Therefore, many scholars are also actively engaged in innovation study on student satisfaction with entrepreneurship courses, for example, Sisilia used the Kano Model to analyze the influencing factors on student satisfaction with entrepreneurship courses and to determine the influence of different indicators on student satisfaction by calculating the coefficient of customer satisfaction (Sisilia and Garmaisa, 2014).…”

Section: Introductionmentioning

confidence: 99%

Influencing Factors of Student Satisfaction With the Teaching Quality of Fundamentals of Entrepreneurship Course Under the Background of Innovation and Entrepreneurship

2021

View full text Add to dashboard Cite

Innovation and entrepreneurship education is a key way to cultivate applied talents in colleges and universities. This article aimed to optimize the teaching quality of the Fundamentals of Entrepreneurship courses by constructing a conceptual model of “teaching quality–student satisfaction.”Teaching quality is divided into four indicator elements, teaching content, teaching methods, teaching conditions, and teaching management. A student satisfaction questionnaire was designed to measure the teaching quality of a Fundamentals of Entrepreneurship course. The reliability and validity of survey data from Linyi University were analyzed using SPSS.20 software. Through correlation and multiple regression analysis, it can be seen that the teaching content, teaching methods, teaching conditions and teaching management have a significant positive correlation with the teaching quality of the Fundamentals of Entrepreneurship course, which is an important factor affecting student’s satisfaction, and there is a certain gap between the expectation of teaching quality and student’s satisfaction. On this basis, suggestions are put forward to improve student satisfaction with the teaching quality of Fundamentals of Entrepreneurship courses, and to provide empirical evidence and recommendations for continuously improving the teaching quality of Fundamentals of Entrepreneurship courses, thereby improving college students’ employment and entrepreneurship ability.

show abstract

Cost-Driven Off-Loading for DNN-Based Applications Over Cloud, Edge, and End Devices

Cited by 137 publications

References 31 publications

HierTrain: Fast Hierarchical Edge AI Learning With Hybrid Parallelism in Mobile-Edge-Cloud Computing

HierTrain: Fast Hierarchical Edge AI Learning With Hybrid Parallelism in Mobile-Edge-Cloud Computing

Energy-Efficient Offloading for DNN-Based Smart IoT Systems in Cloud-Edge Environments

Influencing Factors of Student Satisfaction With the Teaching Quality of Fundamentals of Entrepreneurship Course Under the Background of Innovation and Entrepreneurship

Contact Info

Product

Resources

About