Measure Contribution of Participants in Federated Learning

Wang, Guan; Dang, Charlie Xiaoqian; Zhou, Ziye

doi:10.1109/bigdata47090.2019.9006179

Cited by 158 publications

(99 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the authors of [9], [12], [13] only focused on hFL to share the sample space and did not consider the diversity of services. Moreover, the authors of [11] used Shapley values in vFL to calculate the importance of features, opening the door for investigating hybrid FL and crediting allocation in the context of FL in terms of diversified service types. To the best of our knowledge, device association over RAN slices based on FL is still not considered in current researches.…”

Section: Related Workmentioning

confidence: 99%

“…FL trains nonindependently identically distribution and unbalanced data locally at individual devices and exploits the collaboration of the devices. Specifically, FL is classified into horizontally FL (hFL), vertically FL (vFL), and federated transfer learning based on how data is distributed among various devices in the feature and sample space [10], [11]. Most of the existing related work focuses on hFL to share the sample space or vFL to share the feature space, such as [9], [12], [13].…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we propose an intelligent device association scheme for RAN slicing, called hybrid federated deep reinforcement learning (HDRL) scheme, with the aim to improve network throughput while reducing handoff cost. Considering the large state-action space and the diversity of services, HDRL is designed to consist of two layer model aggregations: 1) Horizontal aggregation: for the same type of services (e.g., eMBB services), we aggregate the parameters of local DRL model on BSs to share the similar samples; 2) Vertical aggregation: for the services of different types (e.g., eMBB and URLLC services), we aggregate the access features of local DRL model on the third encrypted party in vFL [10], where we use Shapley value [11] to evaluate aggregated access features. Numerical results show that in the typical scenarios, our proposed HDRL scheme for device association significantly outperforms the traditional solutions in terms of network throughput and communication efficiency.…”

Section: Introductionmentioning

confidence: 99%

“…(1) We combine DRL and hFL to train distributed data over smart devices, with the aim to retain the privacy of local data. (2) We calculate Shapley values to evaluate the importance of different global access features [10], [11] and promote collaboration between devices. (3) We propose to exploit two levels of aggregation for device association problem.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Device Association for RAN Slicing Based on Hybrid Federated Deep Reinforcement Learning

Liu

Feng

Sun

et al. 2020

IEEE Trans. Veh. Technol.

View full text Add to dashboard Cite

Network slicing (NS) has been widely identified as a key architectural technology for 5G-and-beyond systems by supporting divergent requirements in a sustainable way. In radio access network (RAN) slicing, due to the device-base station (BS)-NS three layer association relationship, device association (including access control and handoff management) becomes an essential yet challenging issue. With the increasing concerns on stringent data security and device privacy, exploiting local resources to solve device association problem while enforcing data security and device privacy becomes attractive. Fortunately, recently emerging federated learning (FL), a distributed learning paradigm with data protection, provides an effective tool to address this type of issues in mobile networks. In this paper, we propose an efficient device association scheme for RAN slicing by exploiting a hybrid FL reinforcement learning (HDRL) framework, with the aim to improve network throughput while reducing handoff cost. In our proposed framework, individual smart devices train a local machine learning model based on local data and then send the model features to the serving BS/encrypted party for aggregation, so as to efficiently reduce bandwidth consumption for learning while enforcing data privacy. Specifically, we use deep reinforcement learning to train the local model on smart devices under a hybrid FL framework, where horizontal FL is employed for parameter aggregation on BS, while vertical FL is employed for NS/BS pair selection aggregation on the encrypted party. Numerical results show that the proposed HDRL scheme can achieve significant performance gain in terms of network throughput and communication efficiency in comparison with some state-of-the-art solutions.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Device Association for RAN Slicing Based on Hybrid Federated Deep Reinforcement Learning

Liu

Feng

Sun

et al. 2020

IEEE Trans. Veh. Technol.

View full text Add to dashboard Cite

show abstract

“…The architecture of the distributed machine learning is shown in Federated Learning is a distributed machine learning approach that enables model training on a large corpus of decentralized data [2]. Federated Machine learning (FML) [34] creates an ecosystem for multiple parties to collaborate on the building models while protecting data privacy for the participants. Instead of transferring data directly into a centralized data warehouse for building machine learning models, Federated Learning allows each party to own the data in its place and still enables all parties to build a machine learning model together.…”

Section: Incentive Mechanism In Federated Learningmentioning

confidence: 99%

A Brief Survey of Data Pricing for Machine Learning

Tang¹,

Lv²,

Wu³

2020

8th International Conference on Signal, Image Processing and Pattern Recognition

View full text Add to dashboard Cite

Big data and machine learning are poised to revolutionize the field of artificial intelligence and represent a step towards building an intelligent society. Big data is considered to be the key to unlocking the next great waves of growth in productivity, the value of data is realized through machine learning. In this survey, we begin with an introduction to the general field of data pricing and distributed machine learning then progress to the main streams of data pricing and mechanism design methods. Our survey will cover several current areas of research within the field of data pricing, including the incentive mechanism design for federated learning, reinforcement learning, auction, crowdsourcing, and blockchain, especially, focus on reward function for machine learning and payment scheme. In parallel, we highlight the pricing scheme in data transactions, focusing on data evaluation via distributed machine learning. To conclude, we discuss some research challenges and future directions of data pricing for machine learning.

show abstract

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

Zheng

Huang

2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

Federated learning (FL) is gradually becoming a key learning paradigm in Privacy-preserving Machine Learning (ML) systems. In FL, a large number of clients cooperate with a central server to learn a shared model without sharing their own data sets.However, since there is a great disparity between the client data sets, standard FL is often hard to tune and suffers from performance degradation due to the inharmony among local models. To this end, in this paper we propose a novel FL scheme, termed client reputation federated learning (CRFL), which dynamically assesses the reputation of the clients participating in FL. Our method leverages techniques from model explanation, and aims at precisely measure each client's impact to the global model. To be specific, we first calculate the saliency-weighted variance on pixelwise relevance scores as the quality factor of a single sample. Then we extract activation function values at the last hidden layer to compute the divergence factor of individual data set. Finally, the server integrates these two factors as an assessment of the client reputation. By leveraging such assessment, CRFL can dynamically adjust the weights of the clients in each aggregation round, thus leading to a significant improvement over the baseline method in terms of model accuracy and

show abstract

Measure Contribution of Participants in Federated Learning

Cited by 158 publications

References 16 publications

Device Association for RAN Slicing Based on Hybrid Federated Deep Reinforcement Learning

Device Association for RAN Slicing Based on Hybrid Federated Deep Reinforcement Learning

A Brief Survey of Data Pricing for Machine Learning

CRFL: A novel federated learning scheme of client reputation assessment via local model inversion

Contact Info

Product

Resources

About