Making resource adaptive to federated learning with COTS mobile devices

Deng, Yongheng; Gu, Shuang; Jiao, Chengbo; Bao, Xing Chuan; Lyu, Feng

doi:10.1007/s12083-021-01284-2

Cited by 3 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most recent research attempts to accelerate the FL process by allocating tasks with different workloads to workers with different capacities to allow heterogeneous workers to finish tasks simultaneously. [24][25][26][27][28][29][30] A joint optimization algorithm is proposed by Chen,27 in which the transmission power of workers and aggregation servers are adjusted to allow transmission efficiency. Instead of tuning the transmission rate to achieve the balance of worker time consumption through tuning the bandwidth, Reference 28 intends to tune the amount of data utilized by different workers so that slow and fast workers can finish the training simultaneously.…”

Section: Related Workmentioning

confidence: 99%

“…These statistics include: (1) F: processor frequency, ( Considering the current literature, each work depends on different system parameters and derives different intermediate results for optimization purposes. However, the current literature always compares their performance with no optimization or random worker selection policy, [20][21][22][23][24][25][26][27][28][29][30] which partially proves the proposed algorithms' necessity. However, it is hard to compare these works against each other because reproducing the experimental results is cumbersome due to the difference between their system setup and programming style.…”

Section: Qualitative Comparisonmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

“…A joint optimization algorithm is proposed by Chen, 27 in which the transmission power of workers and aggregation servers are adjusted to allow transmission efficiency. Instead of tuning the transmission rate to achieve the balance of worker time consumption through tuning the bandwidth, Reference 28 intends to tune the amount of data utilized by different workers so that slow and fast workers can finish the training simultaneously. The Reference 29 proposed to use the dropout method to achieve similar computation time among workers rather than tuning the amount of data.…”

Section: Background and Related Workmentioning

confidence: 99%

See 3 more Smart Citations

FLight: A lightweight federated learning framework in edge and fog computing

Zhu,

Goudarzi,

Buyya

2023

Softw Pract Exp

View full text Add to dashboard Cite

The number of Internet of Things (IoT) applications, especially latency‐sensitive ones, have been significantly increased. So, cloud computing, as one of the main enablers of the IoT that offers centralized services, cannot solely satisfy the requirements of IoT applications. Edge/fog computing, as a distributed computing paradigm, processes, and stores IoT data at the edge of the network, offering low latency, reduced network traffic, and higher bandwidth. The edge/fog resources are often less powerful compared to cloud, and IoT data is dispersed among many geo‐distributed servers. Hence, Federated Learning (FL), which is a machine learning approach that enables multiple distributed servers to collaborate on building models without exchanging the raw data, is well‐suited to edge/fog computing environments, where data privacy is of paramount importance. Besides, to manage different FL tasks on edge/fog computing environments, a lightweight resource management framework is required to manage different incoming FL tasks while does not incur significant overhead on the system. Accordingly, in this article, we propose a lightweight FL framework, called FLight, to be deployed on a diverse range of devices, ranging from resource‐limited edge/fog devices to powerful cloud servers. FLight is implemented based on the FogBus2 framework, which is a containerized distributed resource management framework. Moreover, FLight integrates both synchronous and asynchronous models of FL. Besides, we propose a lightweight heuristic‐based worker selection algorithm to select a suitable set of available workers to participate in the training step to obtain higher training time efficiency. The obtained results demonstrate the efficiency of the FLight. The worker selection technique reduces the training time of reaching 80% accuracy by 34% compared to sequential training, while asynchronous one helps to improve synchronous FL training time by 64%.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Qualitative Comparisonmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

See 2 more Smart Citations

FLight: A lightweight federated learning framework in edge and fog computing

Zhu,

Goudarzi,

Buyya

2023

Softw Pract Exp

View full text Add to dashboard Cite

show abstract

“…Such technological advancement lays a foundation for the emergence of mobile learning platforms, and mobile learning quickly becomes a learning method in style. Now these platforms have covered large populations, their user group expands from on-campus students to nearly everyone in the society, and a huge amount of learning behavior data have accumulated on the platforms during such expansion [8][9][10][11][12][13][14]. The data mining of the mobile learning behavior of learners can help researchers understand the underlying association rules of such behavior and the internal mechanism of the development of thinking during the learning process, thereby giving the true and accurate evaluations on the thought and status of mobile learners [15][16][17][18][19][20].…”

Section: Introductionmentioning

confidence: 99%

Features and Influencing Factors of Mobile Learning Behavior of Employees in Accounting Profession

Wang¹,

Dai²,

Pei³

et al. 2022

Int. J. Emerg. Technol. Learn.

View full text Add to dashboard Cite

Data mining of the mobile learning behavior of learners can help researchers understand the underlying association rules of such behavior and the internal mechanism of the development of their thinking during the learning process, thereby giving the true and accurate evaluations on the thought and status of mobile learners. However, existing models established for mobile learning behavior recognition cannot give satisfactory enough explanations to the changes in the cognition level of learners, so this paper took employees in accounting profession as subjects to analyze the features and influencing factors of their mobile learning behavior. At first, this paper built a neural network model based on the attention mechanism and used it to classify and recognize the input data of mobile learning behavior via original data learning, and feature extraction and analysis. Then, this paper created evaluation scenarios for subjects participating in mobile learning, designed economic management tasks applicable for actual application scenarios, and proposed several evaluation indexes for assessing their learning performance. After that, the explicit learning behavior of the subjects was taken as the criterion for judging whether their learning performance has achieved the desired learning goals or not. At last, the effectiveness of our analysis model was verified by the experimental results.

show abstract