Toward On-Device Federated Learning: A Direct Acyclic Graph-Based Blockchain Approach

Cao, Mingrui; Zhang, Long; Cao, Bin

doi:10.1109/tnnls.2021.3105810

Cited by 62 publications

(25 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…FL as a promising training paradigm, was proposed by Google to tackle the privacy and security problems of centralized machine learning and to alleviate the communication load of the core network [194], [195]. As shown in Fig.…”

Section: Case Studymentioning

confidence: 99%

Blockchain Systems, Technologies, and Applications: A Methodology Perspective

Cao

Wang²,

Zhang

et al. 2023

IEEE Commun. Surv. Tutorials

Self Cite

View full text Add to dashboard Cite

In the past decade, blockchain has shown a promising vision to build trust without any powerful third party in a secure, decentralized and scalable manner. However, due to the wide application and future development from cryptocurrency to the Internet of things, blockchain is an extremely complex system enabling integration with mathematics, computer science, communication and network engineering, etc. By revealing the intrinsic relationship between blockchain and communication, networking and computing from a methodological perspective, it provided a view to the challenge that engineers, experts and researchers hardly fully understand the blockchain process in a systematic view from top to bottom. In this article we first introduce how blockchain works, the research activities and challenges, and illustrate the roadmap involving the classic methodologies with typical blockchain use cases and topics. Second, in blockchain systems, how to adopt stochastic process, game theory, optimization theory, and machine learning to study the blockchain running processes and design the blockchain protocols/algorithms are discussed in details. Moreover, the advantages and limitations using these methods are also summarized as the guide of future work to be further considered. Finally, some remaining problems from technical, commercial and political views are discussed as the open issues. The main findings of this article will provide a survey from a methodological perspective to study theoretical model for blockchain fundamentals understanding, design network service for blockchain-based mechanisms and algorithms, as well as apply blockchain for the Internet of things, etc.

show abstract

Section: Case Studymentioning

confidence: 99%

Blockchain Systems, Technologies, and Applications: A Methodology Perspective

Cao

Wang²,

Zhang

et al. 2023

IEEE Commun. Surv. Tutorials

Self Cite

View full text Add to dashboard Cite

show abstract

“…Pei et al in [10] combined cooperative spectrum sensing with mining in Bitcoin; each secondary user acts both a sensing node for cooperative sensing and a miner in blockchain network. After that, the applications of blockchain to spectrum management were investigated in [11,12], including four typical scenarios, i.e., primary cooperative sharing, secondary cooperative sharing, primary noncooperative sharing, and secondary noncooperative sharing. The applications of blockchain in spectrum management are successively proceeded, some focus on UAV spectrum management [13,14], some focus on the spectrum management in 5G/6G [15][16][17], and some focus on the spectrum management in Internet of Vehicles [18,19].…”

Section: Wireless Communications and Mobile Computingmentioning

confidence: 99%

Blockchain-Empowered High-Frequency Spectrum Management IoT: A Multilayer PBFT Consensus Perspective

Chen

Jian

Qiu³

2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

High frequency (HF) is an important method for long-range communications and even the only mean when satellites are destroyed or interfered, which play an essential role in defense and economic construction. However, noncooperative frequency competition accompanied with power competition results in the continuously deterioration of HF electromagnetic environment. This article endeavors to resolve this issue through proposing blockchain-empowered HF spectrum management. Specifically, massive personal HF devices are organized around the preselected nodes to construct HF spectrum management IoT, further monitor, and share HF data through PBFT (Practical Byzantine Fault Tolerance) protocol. To address the scalability problem during the consensus, a multilayer PBFT consensus protocol is employed. Scalability evaluations show that increasing consensus layers of PBFT greatly reduces the communication complexity. Security assessments illustrate that the security performance will decline with the increase of layers. Tradeoff has been made between the communication complexity and security performance, indicating 2-4 layers PBFT is sufficient, which bring down the communication complexity and also achieve acceptable security performance.

show abstract

“…[15]Federated learning using peer-to-peer network for decentralized orchestration of model weights [16]Peer-to-peer federated learning on graphs [17]Decentralized federated learning for electronic health records [18]Braintorrent: A peer-to-peer environment for decentralized federated learning [13]Lotteryfl: Personalized and communication-efficient federated learning with lottery ticket hypothesis on non-iid datasets [19]Deploy-able privacy preserving collaborative ml Asynchronous [20]Personalized and private peer-to-peer machine learning [21]Personalized cross-silo federated learning on non-iid data [22]Edge-consensus learning: Deep learning on p2p networks with non-homogeneous data [23]Towards on-device federated learning: A direct acyclic graphbased blockchain approach Gossip algorithms, as the name suggests, are a means of P2P communication in distributed systems by leaking information to the neighbors. These are a type of asynchronous algorithms and have been successfully employed in the area of decentralized optimization.…”

Section: Synchronousmentioning

confidence: 99%

“…Here, [21] facilitates learning between clients with similar data and [22] shows that effective deep learning is possible for data residing in different physical locations, even when these data are heterogeneous in nature, and without the need for centralized processing using an edge consensus learning method. [23] introduces a framework for empowering FL using Direct Acyclic Graph (DAG)-based blockchain systematically (DAG-FL).…”

Section: Systemization Of Knowledgementioning

confidence: 99%

“…Other works in the context of FL that define some concrete parameters for measurement of communication cost involve either transmitting or receiving on both devices. Some of these definitions can be found in [28], [23], [34] for further detail. We use a constant scalar α to normalize the cost function and ensure equation ( 3) remains differentiable across all training rounds, which needs to be experimented with and appropriately set if the range of input costs change.…”

Section: B Problem Formulationmentioning

confidence: 99%

See 1 more Smart Citation

TravellingFL: Communication Efficient Peer-to-Peer Federated Learning

Gupta¹,

Luqman²,

Chattopadhyay³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>Machine learning and artificial intelligence are two key emerging technologies in computer science that require vast amounts of data for a meaningful application. The requirement of such a large dataset is usually met by pooling data from various sources, which is often difficult to implement in practice due to strict data privacy constraints. Peer-to-Peer federated learning is a distributed machine learning paradigm with a primary goal of learning a well-performing global model by collaboratively learning a shared model at different data hubs without the need of sharing data. Due to its immense practical applications, there is a growing attention towards various challenges of efficient fed- erated learning including communication efficiency, assumptions on connectivity, data heterogeneity, enhanced privacy, etc. In this paper, we address the communication efficiency of Peer-to-Peer federated learning, modeling it using a graph theoretical frame- work. We show that one can draw from a range of graph-based algorithms to construct an efficient communication algorithm on a connected network, thereby matching the inference efficiency of centralized federated learning as well as that of a consolidated dataset. We conduct experiments with varied graph formations and sizes to validate our claims. </p>

show abstract

Toward On-Device Federated Learning: A Direct Acyclic Graph-Based Blockchain Approach

Cited by 62 publications

References 28 publications

Blockchain Systems, Technologies, and Applications: A Methodology Perspective

Blockchain Systems, Technologies, and Applications: A Methodology Perspective

Blockchain-Empowered High-Frequency Spectrum Management IoT: A Multilayer PBFT Consensus Perspective

TravellingFL: Communication Efficient Peer-to-Peer Federated Learning

Contact Info

Product

Resources

About