A Green Multi-Attribute Client Selection for Over-The-Air Federated Learning: A Grey-Wolf-Optimizer Approach DOI Open Access
Maryam Ben Driss, Essaïd Sabir, Halima Elbiaze

и другие.

ACM Transactions on Modeling and Performance Evaluation of Computing Systems, Год журнала: 2024, Номер unknown

Опубликована: Дек. 6, 2024

Federated Learning (FL) has gained attention across various industries for its capability to train machine learning models without centralizing sensitive data. While this approach offers significant benefits such as privacy preservation and decreased communication overhead, it presents several challenges, including deployment complexity interoperability issues, particularly in heterogeneous scenarios or resource-constrained environments. Over-the-air (OTA) FL was introduced tackle these challenges by disseminating model updates necessitating direct device-to-device connections centralized servers. However, OTA-FL brought forth limitations associated with heightened energy consumption network latency. In paper, we propose a multi-attribute client selection framework employing the grey wolf optimizer (GWO) strategically control number of participants each round optimize process while considering accuracy, energy, delay, reliability, fairness constraints participating devices. We evaluate performance our terms loss minimization, convergence time reduction, efficiency. experimental evaluation, assessed compared against existing state-of-the-art methods. Our results demonstrate that proposed GWO-based outperforms baselines metrics. Specifically, achieves notable reduction loss, accelerates time, enhances efficiency maintaining high reliability indicators.

Язык: Английский

Data and Model Poisoning Backdoor Attacks on Wireless Federated Learning, and the Defense Mechanisms: A Comprehensive Survey DOI
Yichen Wan, Youyang Qu, Wei Ni

и другие.

IEEE Communications Surveys & Tutorials, Год журнала: 2024, Номер 26(3), С. 1861 - 1897

Опубликована: Янв. 1, 2024

Due to the greatly improved capabilities of devices, massive data, and increasing concern about data privacy, Federated Learning (FL) has been increasingly considered for applications wireless communication networks (WCNs). Wireless FL (WFL) is a distributed method training global deep learning model in which large number participants each train local on their datasets then upload updates central server. However, general, nonindependent identically (non-IID) WCNs raises concerns robustness, as malicious participant could potentially inject "backdoor" into by uploading poisoned or models over WCN. This cause misclassify inputs specific target class while behaving normally with benign inputs. survey provides comprehensive review latest backdoor attacks defense mechanisms. It classifies them according targets (data poisoning poisoning), attack phase (local collection, training, aggregation), stage before aggregation, during after aggregation). The strengths limitations existing strategies mechanisms are analyzed detail. Comparisons methods designs carried out, pointing noteworthy findings, open challenges, potential future research directions related security privacy WFL.

Язык: Английский

Процитировано

19
Over-the-Air Federated Learning and Optimization DOI
Jingyang Zhu, Yuanming Shi, Yong Zhou

и другие.

IEEE Internet of Things Journal, Год журнала: 2024, Номер 11(10), С. 16996 - 17020

Опубликована: Янв. 10, 2024

Federated learning (FL), as an emerging distributed machine paradigm, allows a mass of edge devices to collaboratively train global model while preserving privacy. In this tutorial, we focus on FL via over-the-air computation (AirComp), which is proposed reduce the communication overhead for over wireless networks at cost compromising in performance due aggregation error arising from channel fading and noise. We first provide comprehensive study convergence AirComp-based FEDAVG (AIRFEDAVG) algorithms under both strongly convex non-convex settings with constant diminishing rates presence data heterogeneity. Through asymptotic analysis, characterize impact bound insights system design guarantees. Then derive AIRFEDAVG objectives. For different types local updates that can be transmitted by (i.e., model, gradient, difference), reveal transmitting may cause divergence training procedure. addition, consider more practical signal processing schemes improve efficiency further extend analysis forms caused these schemes. Extensive simulation results objective functions, information, verify theoretical conclusions.

Язык: Английский

Процитировано

4
6G Wireless Communications and Artificial Intelligence-Controlled Reconfigurable Intelligent Surfaces: From Supervised to Federated Learning DOI Creative Commons

Evangelos A. Zaoutis,

G. Liodakis,

Anargyros T. Baklezos

и другие.

Applied Sciences, Год журнала: 2025, Номер 15(6), С. 3252 - 3252

Опубликована: Март 17, 2025

The new generation of wireless communication technologies is already in development. Sixth Generation (6G) mobile communications are designed to push the limits for more bandwidth, connected devices with minimal power requirements, and better signal quality. Previous used Fifth (5G) inadequate handle requirements alone. One proposed solutions use Reconfigurable Intelligent Surfaces (RISs). These surfaces, when combined Artificial Intelligence (AI), may be a very powerful means achieving this. In this paper, we review studies that focus on RISs controlled by AI determining concept Smart Radio Environment (SRE) 6G networks. We examine applications span from Supervised Federated Learning (FL) as enabled rise Edge Computing. As expected have enhanced capabilities perform computing locally, thus reducing need transfer data central hub, opportunities created extensive FL. context, research FL RIS-aided SRE.

Язык: Английский

Процитировано

0
Seizure type classification algorithm based on multi-dimensional brain network feature selection DOI Creative Commons
Duanpo Wu, Yixiao Mao, Yuhan Gao

и другие.

Fundamental Research, Год журнала: 2025, Номер unknown

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
A Survey on Federated Learning for Reconfigurable Intelligent Metasurfaces-Aided Wireless Networks DOI Creative Commons
S. Das, Benoı̂t Champagne, Ioannis N. Psaromiligkos

и другие.

IEEE Open Journal of the Communications Society, Год журнала: 2024, Номер 5, С. 1846 - 1879

Опубликована: Янв. 1, 2024

Wireless networks are increasingly relying on machine learning (ML) paradigms to provide various services at the user level. Yet, it remains impractical for users offload their collected data set a cloud server centrally training local ML model. Federated (FL), which aims collaboratively train global model by leveraging distributed wireless computation resources across without exchanging information, is therefore deemed as promising solution enabling intelligent in data-driven society of future. Recently, reconfigurable metasurfaces (RIMs) have emerged revolutionary technology, offering controllable means increasing signal diversity and reshaping transmission channels, implementation constraints traditionally associated with multi-antenna systems. In this paper, we present comprehensive survey recent works applications FL RIM-aided communications. We first review fundamental basis an emphasis mechanisms, well operating principles RIMs, including tuning operation modes, deployment options. then proceed in-depth literature FL-based approaches recently proposed three key interrelated problems networks, namely: channel estimation (CE), passive beamforming (PBF) resource allocation (RA). each case, illustrate discussion introducing expanded (EFL) framework only subset active partake process, thereby allowing reduce overhead. Lastly, discuss some current challenges research avenues full potential future extremely large-scale multiple-input-multiple-output (XL-MIMO) networks.

Язык: Английский

Процитировано

2
Unleashing Edgeless Federated Learning With Analog Transmissions DOI
Howard H. Yang, Zihan Chen, Tony Q. S. Quek

и другие.

IEEE Transactions on Signal Processing, Год журнала: 2024, Номер 72, С. 774 - 791

Опубликована: Янв. 1, 2024

We demonstrate that merely analog transmissions and match filtering can realize the function of an edge server in federated learning (FL). Therefore, a network with massively distributed user equipments (UEs) achieve large-scale FL without server. also develop training algorithm allows UEs to continuously perform local computing being interrupted by global parameter uploading, which exploits full potential UEs' processing power. derive convergence rates for proposed schemes quantify their efficiency. The analyses reveal when interference obeys Gaussian distribution, retrieves rate server-based FL. But if distribution is heavy-tailed, then heavier tail, slower converges. Nonetheless, system run time be largely reduced enabling computation parallel communication, whereas gain particularly pronounced communication latency high. These findings are corroborated via extensive simulations.

Язык: Английский

Процитировано

0
GWO-Boosted Multi-Attribute Client Selection for Over- The-Air Federated Learning DOI
Maryam Ben Driss, Essaïd Sabir, Halima Elbiaze

и другие.

Опубликована: Май 6, 2024

Federated Learning (FL) has gained popularity across various industries due to its ability train machine learning models without explicit sharing of sensitive data. While this paradigm offers significant advantages such as privacy preser-vation and reduced communication overhead, it also comes with several challenges deployment complexity interoperability issues, particularly in heterogeneous scenarios or resource-constrained environments. Over-the-air (OTA) FL was introduced address those by model updates the need for direct device- to-device connections cen-tralized servers. However, OTA - induces some issues related increased energy consumption, wireless channel variability, network latency. In paper, we propose a multi-attribute client selection framework using Grey Wolf optimizer limit number participants each round optimize process while considering energy, delay, reliability, fairness constraints participating devices. We analyze performance our approach terms loss, convergence time, overall accuracy. Our experimental results show that proposed can lower consumption up 43% compared random method.

Язык: Английский

Процитировано

0
A Multifaceted Survey on Federated Learning: Fundamentals, Paradigm Shifts, Practical Issues, Recent Developments, Partnerships, Trade-Offs, Trustworthiness, and Ways Forward DOI Creative Commons
Abdul Majeed, Seong Oun Hwang

IEEE Access, Год журнала: 2024, Номер 12, С. 84643 - 84679

Опубликована: Янв. 1, 2024

Язык: Английский

Процитировано

0
Federated learning for millimeter-wave spectrum in 6G networks: applications, challenges, way forward and open research issues DOI Creative Commons
Faizan Qamar, Syed Hussain Ali Kazmi, Maraj Uddin Ahmed Siddiqui

и другие.

PeerJ Computer Science, Год журнала: 2024, Номер 10, С. e2360 - e2360

Опубликована: Окт. 9, 2024

The emergence of 6G networks promises ultra-high data rates and unprecedented connectivity. However, the effective utilization millimeter-wave (mmWave) as a critical enabler foreseen potential in 6G, poses significant challenges due to its unique propagation characteristics security concerns. Deep learning (DL)/machine (ML) based approaches emerged solutions; however, DL/ML contains centralization privacy issues. Therefore, federated (FL), an innovative decentralized paradigm, offers promising avenue tackle these by enabling collaborative model training across distributed devices while preserving privacy. After comprehensive exploration FL enabled networks, this review identifies specific applications mmWave communications context networks. Thereby, article discusses particular faced adaption communication 6G; including bandwidth consumption, power consumption synchronization requirements. In view identified challenges, study proposed way forward called Federated Energy-Aware Dynamic Synchronization with Bandwidth-Optimization (FEADSBO). Moreover, highlights pertinent open research issues synthesizing current advancements efforts. Through review, we provide roadmap harness synergies between mmWave, offering insights reshape landscape

Язык: Английский

Процитировано

0
A Green Multi-Attribute Client Selection for Over-The-Air Federated Learning: A Grey-Wolf-Optimizer Approach DOI Open Access
Maryam Ben Driss, Essaïd Sabir, Halima Elbiaze

и другие.

ACM Transactions on Modeling and Performance Evaluation of Computing Systems, Год журнала: 2024, Номер unknown

Опубликована: Дек. 6, 2024

Federated Learning (FL) has gained attention across various industries for its capability to train machine learning models without centralizing sensitive data. While this approach offers significant benefits such as privacy preservation and decreased communication overhead, it presents several challenges, including deployment complexity interoperability issues, particularly in heterogeneous scenarios or resource-constrained environments. Over-the-air (OTA) FL was introduced tackle these challenges by disseminating model updates necessitating direct device-to-device connections centralized servers. However, OTA-FL brought forth limitations associated with heightened energy consumption network latency. In paper, we propose a multi-attribute client selection framework employing the grey wolf optimizer (GWO) strategically control number of participants each round optimize process while considering accuracy, energy, delay, reliability, fairness constraints participating devices. We evaluate performance our terms loss minimization, convergence time reduction, efficiency. experimental evaluation, assessed compared against existing state-of-the-art methods. Our results demonstrate that proposed GWO-based outperforms baselines metrics. Specifically, achieves notable reduction loss, accelerates time, enhances efficiency maintaining high reliability indicators.

Язык: Английский

Процитировано

0