区块链赋能的设备端联邦学习

版权申诉

5星 · 超过95%的资源 194 浏览量更新于2024-08-19 收藏 2MB PDF 举报

"Blockchained On-Device Federated Learning——一种基于区块链的设备端联邦学习架构" 在当前的数字化时代，数据安全与隐私保护已经成为至关重要的议题。联邦学习（Federated Learning, FL）作为一种分布式机器学习方法，允许设备在不共享本地数据的情况下协同训练模型，从而兼顾了学习性能和用户隐私。然而，传统的联邦学习架构通常依赖于中心服务器来协调各个设备的模型更新，这不仅可能成为系统的瓶颈，还存在单点故障的风险。 "Blockchained On-Device Federated Learning" 提出了一种创新的解决方案，即通过区块链技术替代中央服务器的角色。这里的区块链不仅提供了去中心化的数据交换平台，还引入了验证机制，确保每个设备上传的模型更新是可靠且不可篡改的。这种机制进一步强化了联邦学习的去中心化特性，并为参与者提供了激励，促进了系统的积极性和公正性。文章深入研究了BlockFL的端到端学习延迟问题。在联邦学习中，延迟因素包括通信延迟、计算延迟以及区块链中的共识延迟。通过分析这些延迟组件，作者们建立了一个完整的延迟模型，并探索了最优的区块生成速率，以最小化总的延迟时间。优化区块生成速率是关键，因为它直接影响到区块链网络的吞吐量和响应速度。关键词中的"联邦学习"强调了在分布式环境中进行模型训练的能力，而"区块链"则突显了系统的核心技术创新，即使用区块链技术来促进模型更新的安全交换和验证。这两者的结合，为解决传统联邦学习的中心化问题提供了新的视角。在介绍部分，作者指出未来的无线系统需要实现低延迟和高可靠性。而设备端机器学习（On-device Machine Learning）正是满足这一需求的一种策略，它允许设备在无连接时也能做出决策。然而，仅依靠本地数据训练的模型可能无法达到理想的效果，因此需要更多的数据样本，这正是联邦学习的价值所在。通过BlockFL，设备可以利用区块链的共识机制，在无需集中式数据或协调的情况下进行协作学习，从而提高整体模型的性能，同时保护用户隐私。 "Blockchained On-Device Federated Learning" 是一篇探讨如何结合区块链技术和联邦学习以优化设备端学习性能、减少延迟并增强安全性的研究论文。它不仅提供了理论分析，还为实际应用中的系统优化提供了指导，对于理解区块链在机器学习领域的潜在应用具有重要意义。

Blockchained On-Device Federated Learning

Hyesung Kim, Jihong Park

†

, Mehdi Bennis

†

, and Seong-Lyun Kim

Abstract—By leveraging blockchain, this letter proposes a

blockchained federated learning (BlockFL) architecture where

local learning model updates are exchanged and veriﬁed. This

enables on-device machine learning without any centralized

training data or coordination by utilizing a consensus mechanism

in blockchain. Moreover, we analyze an end-to-end latency model

of BlockFL and characterize the optimal block generation rate by

considering communication, computation, and consensus delays.

Index Terms—On-device machine learning, federated learning,

blockchain, latency.

I. INTRODUCTION

UTURE wireless systems are envisaged to ensure low

latency and high reliability anywhere and anytime [1]–

[3]. To this end, on-device machine learning is a compelling

solution wherein each device stores a high-quality machine

learning model and is thereby capable of make decisions, even

when it loses connectivity. Training such an on-device machine

learning model requires more data samples than each device’s

local samples, and necessitates sample exchanges with other

devices [4]–[6]. In this letter, we tackle the problem of training

each device’s local model by federating with other devices.

One key challenge is that local data samples are owned

by each device. Thus, the exchanges should keep the raw

data samples private from other devices. For this purpose, as

proposed in Google’s federated learning (FL) [4], [5], referred

to as vanilla FL, each device exchanges its local model update,

i.e., learning model’s weight and gradient parameters, from

which the raw data cannot be derived. As illustrated in Fig. 1-

a, the vanilla FL’s exchange is enabled by the aid of a central

server that aggregates and takes an ensemble average of all

the local model updates, yielding a global model update.

Then, each device downloads the global model update, and

computes its next local update until the global model training

is completed [5]. Due to these exchanges, the vanilla FL’s

training completion latency might be tens of minutes or more,

as demonstrated in Google’s keyboard application [7].

The limitation of the vanilla FL operation is two-fold.

Firstly, it relies on a single central server, which is vulnerable

to the server’s malfunction. This incurs inaccurate global

model updates distorting all local model updates. Secondly,

it does not reward the local devices. A device having a larger

number of data samples contributes more to the global training.

This work was partly supported by Institute of Information & communications

Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT)

(No.2018-0-00170, Virtual Presence in Moving Objects through 5G), Basic Science

Research Foundation of Korea(NRF) grant funded by the Ministry of Science and ICT

(NRF-2017R1A2A2A05069810), and the Mobile Edge Intelligence at Scale (ELLIS)

project at the University of Oulu.

H. Kim and S.-L. Kim are with School of Electrical and Electronic Engineering,

Yonsei University, Seoul, Korea (email: {hskim, slkim}@ramo.yonsei.ac.kr).

†

J. Park and

†

M. Bennis are with the Centre for Wireless Communications, University

of Oulu, 4500 Oulu, Finland (email: {jihong.park, mehdi.bennis}@oulu.ﬁ).

Federated

learning

1

block propagation

ledger

data reward

Blockchain

<latexit sha1_base64="djQ//6SAWZE9g9pLD4zKXkY4NsY=">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</latexit>

<latexit sha1_base64="WzkMiyqvXvKBkIcCFU2ODIBMPrM=">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</latexit>

<latexit sha1_base64="ApOBfEbRi+xNjvTFkJBWkt01Odc=">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</latexit>

<latexit sha1_base64="K3VThbu2uLljlyKEFsfCndz/nyE=">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</latexit>

mining reward

block generation

cross

veriﬁcation

(1)

(2)

(3)

(4)

(a) Vanilla FL. (b) Proposed BlockFL.

global model

local data

local model

Fig. 1. An illustration of (a) the vanilla federated learning (FL) [4], [5] and

(b) the proposed blockchained FL (BlockFL) architectures.

Without providing compensation, such a device is less willing

to federate with the other devices possessing few data samples.

In order to resolve these pressing issues, by leveraging

blockchain [8], [9] in lieu of the central server, we propose a

blockchained FL (BlockFL) architecture, where the blockchain

network enables exchanging devices’ local model updates

while verifying and providing their corresponding rewards.

BlockFL overcomes the single point of failure problem and

extends the range of its federation to untrustworthy devices in

a public network thanks to a validation process of the local

training results. Moreover, by providing rewards proportional

to the training sample sizes, BlockFL promotes the federation

of more devices with a larger number of training samples.

As shown in Fig. 1-b, the logical structure of BlockFL

consists of devices and miners. The miners can physically be

either randomly selected devices or separate nodes such as

network edges (i.e., base stations in cellular networks), which

are relatively free from energy constraints in mining process.

The operation of BlockFL is summarized as follows: Each

device computes and uploads the local model update to its

associated miner in the blockchain network; Miners exchange

and verify all the local model updates, and then run the Proof-

of-Work (PoW) [8]; Once a miner completes the PoW, it

generates a block where the veriﬁed local model updates are

recorded; and ﬁnally, the generated block storing the aggregate

local model updates is added to a blockchain, also known as

distributed ledger, and is downloaded by devices. Each device

computes the global model update from the new block.

Note that the global model update of BlockFL is computed

locally at each device. A miner’s or a device’s malfunction

does not affect other devices’ global model updates. For the

sake of these beneﬁts, in contrast to the vanilla FL, BlockFL

needs to account for the extra delay incurred by the blockchain

network. To address this, the end-to-end latency model of

BlockFL is formulated by considering communication, compu-

tation, and the PoW delays. The resulting latency is minimized

by adjusting the block generation rate, i.e., the PoW difﬁculty.

arXiv:1808.03949v2 [cs.IT] 1 Jul 2019

下载后可阅读完整内容，剩余4页未读，立即下载

万千之喜

粉丝: 36

区块链赋能的设备端联邦学习

A blockchain-orchestrated Federated Learning.pdf

Advances and Open Problems in Federated Learning.pdf

python2-websocket-client-0.54.0-1.el7.noarch.rpm

Blockchained On-Device Federated Learning

Privacy-Preserving Machine Learning Using Federated Learning and Secure Aggregation

tff.learning.build_federated_averaging_process与tff.learning.build_weighted_fed_avg区别

Multi-objective Evolutionary Federated Learning

请你提供5篇最新的federated learning论文

ubuntu22.04安装mysql-5.6.16.tar.gz

出现module 'tensorflow_federated.python.learning' has no attribute 'optimizers'错误是什么原因

最新资源