Decentralized Federated Learning: Fundamentals, State of the Art, Frameworks, Trends, and Challenges
Open accessThe DOI matches its Crossref record; 230 of the 230 references it cites were found in a registry or an index.
Abstract
In recent years, Federated Learning (FL) has gained relevance in training collaborative models without sharing sensitive data. Since its birth, Centralized FL (CFL) has been the most common approach in the literature, where a central entity creates a global model. However, a centralized approach leads to increased latency due to bottlenecks, heightened vulnerability to system failures, and trustworthiness concerns affecting the entity responsible for the global model creation. Decentralized Federated Learning (DFL) emerged to address these concerns by promoting decentralized model aggregation and minimizing reliance on centralized architectures. However, despite the work done in DFL, the literature has not (i) studied the main aspects differentiating DFL and CFL; (ii) analyzed DFL frameworks to create and evaluate new solutions; and (iii) reviewed application scenarios using DFL. Thus, this article identifies and analyzes the main fundamentals of DFL in terms of federation architectures, topologies, communication mechanisms, security approaches, and key performance indicators. Additionally, the paper at hand explores existing mechanisms to optimize critical DFL fundamentals. Then, the most relevant features of the current DFL frameworks are reviewed and compared. After that, it analyzes the most used DFL application scenarios, identifying solutions based on the fundamentals and frameworks previously defined. Finally, the evolution of existing DFL solutions is studied to provide a list of trends, lessons learned, and open challenges.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Data breaches within third-party services could lead to the loss of proprietary vehicle models or private data, compromising user privacy and vehicle security on a large scale [7, 12, 22].”
From Privacy-preserving Machine Learning in Internet of Vehicle Applications: Fundamentals, Recent Advances, and Future Direction (2025) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In DFL, achieving consensus among participating nodes is crucial to ensure that model updates are coherent and reflect the contributions of all peers [12].”
From Privacy-preserving Machine Learning in Internet of Vehicle Applications: Fundamentals, Recent Advances, and Future Direction (2025) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Data security is another critical challenge, as the transmission and storage of sensitive data create vulnerabilities that may lead to breaches, exposing both individual and aggregate information [7, 12, 22].”
From Privacy-preserving Machine Learning in Internet of Vehicle Applications: Fundamentals, Recent Advances, and Future Direction (2025) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The choice of topology affects the overall privacy, robustness, flexibility, fault tolerance, and communication costs of the network [12].”
From Privacy-preserving Machine Learning in Internet of Vehicle Applications: Fundamentals, Recent Advances, and Future Direction (2025) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“While CFL is effective in various scenarios, it has limitations such as a single point of failure, bottlenecks at the server, and various privacy risks associated with centralized data aggregation [12].”
From Privacy-preserving Machine Learning in Internet of Vehicle Applications: Fundamentals, Recent Advances, and Future Direction (2025) · cites this paper as [12]
Experimental — accuracy not yet measured
4 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“[16] 2022 Privacy and protection of medical data exchanged between healthcare entities Cross-silo Decentralized Real ResNet18 ResNet50 DenseNet121 Custom algorithm Synchronous Sequential ImageNet [166] Private”
BackgroundLearning Multiple Layers of Features from Tiny Images
Krizhevsky, A. (2024) · doi:10.57702/zp44cu3g · Cited by 23,395
Are we ready for autonomous driving? The KITTI vision benchmark suite
Geiger, A., Lenz, P., & Urtasun, R. (2012) · 2012 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2012.6248074 · Cited by 14,897 · cited 3×
“[22] 2021 Novel method proposed by DFL with P2P communication between autonomous vehicles Cross-device Centralized Decentralized Simulation CNN Custom algorithm Synchronous Schematic MNIST [172] KITTI [219]”
MethodInfluentialCommunication-Efficient Learning of Deep Networks from Decentralized Data
McMahan, H. B., Moore, E., Ramage, D., Hampson, S., & Arcas, B. A. Y. (2016) · arXiv · doi:10.48550/arxiv.1602.05629 · Cited by 14,215 · cited 5×
“Furthermore, the controller can select participants based on a preassigned probability calculated according to participation in the federation [4].”
BackgroundInfluentialDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204 · cited 1×
“Specifically, it is a procedure to train a model using a second NN as a reference instead of learning directly from the ground truth of the data [141].”
MethodnuScenes: A Multimodal Dataset for Autonomous Driving
Caesar, H., Bankiti, V., Lang, A. H., Vora, S., Liong, V. E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., & Beijbom, O. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01164 · Cited by 7,152
Learning Word Vectors for Sentiment Analysis
Maas, A. L., Daly, R. E., Pham, P. T., Huang, D., Ng, A. Y., & Potts, C. (2011) · Cited by 6,961
The multimodal brain tumor image segmentation benchmark (BRATS)
Menze, B. H., Jakab, A., Bauer, S., Kalpathy-Cramer, J., Farahani, K., Kirby, J., Burren, Y., Porz, N., Slotboom, J., Wiest, R., et al. (2015) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2014.2377694 · Cited by 6,781 · cited 3×
“[96] 2022 Detect brain tumor segmentation, using representative clinical datasets Cross-device Centralized Decentralized Real U-Net CFA FedPer Synchronous Asynchronous Random BRATS [167] Private [88] 2022 Detect misconduct model on specific sites from any learning iteration Cross-device Centralized Decentralized Simul…”
MethodInfluentialFederated learning: Challenges, methods, and future directions
Li, T., Sahu, A. K., Talwalkar, A., & Smith, V. (2020) · IEEE Signal Processing Magazine · doi:10.1109/msp.2020.2975749 · Cited by 6,210 · cited 1×
“Regarding CFL, several works have attracted immense interest from academia and industry [37].”
BackgroundFashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
Xiao, H., Rasul, K., & Vollgraf, R. (2017) · arXiv · doi:10.48550/arxiv.1708.07747 · Cited by 6,108 · cited 2×
“[222] 2023 Blockchain-enhanced DFL for efficient and privacy-protective data sharing Cross-device Decentralized Simulation CNN FedAvg Asynchronous Scheduling FMNIST [190]”
Method
Research works printed but not linked: 13
- 2022, DISCO—Distributed collaborative machine learning
- Li, 2022, Mining latent relationships among clients: Peer-to-peer federated learning with adaptive neighbor matching, arXiv:2203.12285 · arXiv
- 2022, Tensorflow federated
- Go, 2009, Twitter sentiment classification using distant supervision
- 2016, Geo dataset browser
- Dos, 2021, Predicting human eye diseases
- Savazzi, 2020, Federated learning: mmWave MIMO radar dataset for testing
- 10.1145/3340531.3411866 · DOI
- 2016, Avito context ad clicks
- 2021, IoT device identification
- Koch, 2021, On digital ethics for artificial intelligence in hybrid military operations
- Reiser, 2022, Creiser/drone-detection
- Fisher, 2019, CVONLINE: Image databases
Cited by
Items 101–200 of 626
Guo, F., Li, S., Yang, H., Dong, C., Chen, Y., & Li, G. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3507183 · Cited by 10
“However, ensuring the complete reliability of the central server in practical scenarios is challenging, potentially exposing user privacy to risks [13].”
AI-based robots in industrialized building manufacturing
Wang, M., Cai, J., Hu, D., Hu, Y., Han, Z., & Li, S. (2025) · Frontiers of Engineering Management · doi:10.1007/s42524-025-4099-x · Cited by 11
“For instance, federated learning allows robots to aggregate decentralized learning experiences, enabling them to benefit from collective insights without compromising data privacy (Martínez Beltrán et al., 2023).”
Asynchronous Decentralized Federated Anomaly Detection for 6G Networks
Liu, Y., & Yang, K. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3538020 · Cited by 12
“These issues can also result in incomplete participation in global updates, ultimately degrading the performance of a federated learning framework [12].”
SRFL: A Swarm-Reputation-Based Autonomic Federated Learning Framework for AIoT
Zhang, W., Du, M., Guo, X., & Xiong, N. N. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3509264 · Cited by 16
“Federated learning (FL) [8], [9], [10], an innovative decentralized learning paradigm, enables model collaborative training across multiple entities without exchanging local data, thereby significantly mitigating the risk of privacy breaches.”
Disaster identification scheme based on federated learning and Cognitive Internet of Vehicles
Anjum, M. J., Farooq, M. S., Umer, T., & Shaheen, M. (2025) · Computer Communications · doi:10.1016/j.comcom.2025.108216 · Cited by 9
Tricomi, G., Cicceri, G., Ficili, I., Vitabile, S., Merlino, G., & Puliafito, A. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107991 · Cited by 9
A Lightweight Decentralized Federated Learning Framework for the Industrial Internet of Things
Yang, C., Wang, J., Liu, Y., Ding, Y., Liu, Z., & Wang, S. (2025) · Ad Hoc Networks · doi:10.1016/j.adhoc.2025.103962 · Cited by 11
The governance of federated learning: a decision framework for organisational archetypes
Barbereau, T., Fernández, J. D., & Menci, S. P. (2025) · Data & Policy · doi:10.1017/dap.2025.10020 · Cited by 15
“By removing the dependency on a central coordination agent and distributing responsibility among clients, decentralised FL fosters a more resilient and efficient collaboration (Martínez Beltrán et al., 2023).”
Shrestha, A. K., Khan, F. A., Shaikh, M. A., Jaberzadeh, A., & Geng, J. (2023) · IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) · doi:10.1109/uemcon59035.2023.10316006 · Cited by 10
“Decentralized machine learning, particularly Semi-Decentralized Federated Learning (SDFL), offers a promising solution to these issues by distributing model training across multiple devices, allowing data to remain on-device and preserving privacy while enabling collaborative learning [2], [3].”
BackgroundMethodFRACTAL: Data-Aware Clustering and Communication Optimization for Decentralized Federated Learning
Ma, Q., Liu, J., Xu, H., Jia, Q., & Xie, R. (2024) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2024.3403381 · Cited by 12
“Fullyconnectedtopologies[15],[16]allowworkerstotrans-mittheirmodelstoeachotherthroughpeer-to-peer(P2P) communication.”
A decentralised federated learning scheme for heterogeneous devices in cognitive IoT
Ge, H., Yang, X., Wang, J., & Lv, Z. (2024) · International Journal of Cognitive Computing in Engineering · doi:10.1016/j.ijcce.2024.08.001 · Cited by 9
Hybridized Shield: A Framework for Backdoor Detection in Secure Federated Learning Systems
Rahman, G., Saeed-uz-zaman, Li, B., & Muzamal, J. H. (2024) · IEEE International Conference on Big Data and Artificial Intelligence (BDAI) · doi:10.1109/bdai62182.2024.10692923 · Cited by 9
Chen, Q., Cheng, H., Liang, Y., Zhu, G., Li, M., & Jiang, H. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3499375 · Cited by 11
“W ITH the explosive development of Internet of Things (IoT) technologies, massive amounts of data have been generated at the edge side [1].”
Khan, A. N., Khan, Q. W., Rizwan, A., Ahmad, R., & Kim, D. (2024) · Internet of Things · doi:10.1016/j.iot.2024.101476 · Cited by 15
Hu, B., Zhong, R., Feng, Y., Yang, J., Li, P., Chen, M., & Tan, J. (2025) · International Journal of Production Research · doi:10.1080/00207543.2024.2448604 · Cited by 10
Truong, V. V., Quan, P. K., Park, D.-H., & Kim, T. (2025) · IEEE Access · doi:10.1109/access.2025.3542772 · Cited by 9
“In addition to the impact of topologies, existing studies on [9], [18], [19], [20], [21], and [22] are particularly limited in terms of the effects of heterogeneous computing resources and communication bandwidth.”
Rahnama, S. M. (2025) · IEEE Access · doi:10.1109/access.2025.3568842 · Cited by 17
Healthcare federated learning: a survey of applications and frameworks
Pais, V., Rao, S., & Muniyal, B. (2025) · International Journal of Computers and Applications · doi:10.1080/1206212x.2025.2496913 · Cited by 19
“In a decentralized approach, clients communicate and update models directly between themselves without a central coordinating server [5].”
Malpetti, D., Scutari, M., Gualdi, F., van Setten, J., van der Laan, S. W., Haitjema, S., Lee, A. M., Hering, I., & Mangili, F. (2025) · Frontiers in Digital Health · doi:10.3389/fdgth.2025.1644291 · Cited by 13
Khan, A. A., Timilsina, L., Muriithi, G., Arsalan, A., Moghassemi, A., Papari, B., Özkan, G., Edrington, C. S., Boghrabadi, N. S., & Wang, Z. (2025) · IEEE Access · doi:10.1109/access.2025.3613231 · Cited by 14
Explainable federated transformer framework for joint leukemia classification and stage prediction
Parwez, K., Sohail, S. I., Akram, A., Rashid, J., Atteia, G. E., & Sarwar, N. (2026) · Scientific Reports · doi:10.1038/s41598-025-34715-8 · Cited by 12
“A parallel line of work in federated learning offers promising strategies for decentralized model training across multiple clients without exchanging raw data 21 .”
Ahad, A., Ahmed, K. I., Ullah, F., Sheikh, M. A., Tahir, M., Hayajneh, M., & Pires, I. M. (2026) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.70065 · Cited by 15
Energy-Efficient Connectivity-Aware Learning Over Time-Varying D2D Networks
Parasnis, R., Hosseinalipour, S., Chu, Y.-W., Chiang, M., & Brinton, C. G. (2024) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2024.3374591 · Cited by 9
“This had led to varying degrees of decentralization in FL, reaching fully decentralized, serverless settings that sit at the opposite extreme of the star topology [4], [5], [6], [7], [8], [9], [10].”
BackgroundYan, Z., & Li, D. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3400512 · Cited by 9
“To address these limitations of CFL, DFL as a FEEL framework without the central server, is proposed [5].”
BackgroundMethodInfluentialMajeed, A., & Hwang, S. O. (2024) · IEEE Access · doi:10.1109/access.2024.3413069 · Cited by 14
The Impact of Federated Learning on Improving the IoT-Based Network in a Sustainable Smart Cities
Naeem, M. A., Meng, Y., & Chaudhary, S. (2024) · Electronics · doi:10.3390/electronics13183653 · Cited by 10
Rahmani, A. M., Alsubai, S., Alanazi, A., Alqahtani, A., Zaidi, M. M., & Hosseinzadeh, M. (2024) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2024.109812 · Cited by 9
Mobility-Aware Decentralized Federated Learning for Autonomous Underwater Vehicles
He, H., Du, J., Jiang, C., Wang, J., Song, J., & Han, Z. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3557803 · Cited by 7
“Despite the advantages, most FL approaches still required the central server to aggregate and process the model-related parameters from the IoT devices, thereby introducing potential bottlenecks and single-point failure risk [10], [11].”
Zero-Trust Federated Learning via 6G URLLC for Vehicular Communications
Asad, M., Otoum, S., & Ouni, B. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3560009 · Cited by 8
“…conceived to optimize user experiences across multiple devices, FL has since found broad applications, including in healthcare, finance, and particularly in vehicular networks, providing real-time analytics on traffic conditions and environmental hazards without compromising data privacy [15].”
Sequential Federated Learning in Hierarchical Architecture on Non-IID Datasets
Yan, X., Zuo, S., Fan, R., Hu, H., Shen, L., Zhao, P., & Luo, Y. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3573928 · Cited by 14
“In a real federated learning (FL) system, communication overhead for passing model parameters between the clients and the parameter server (PS) is often a bottleneck.”
DRACO: Decentralized Asynchronous Federated Learning Over Row-Stochastic Wireless Networks
Jeong, E., & Kountouris, M. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3574098 · Cited by 12
“While both decentralization and asynchrony offer significant advantages in flexibility and resource efficiency, their combination introduces additional challenges in achieving global consensus without incurring substantial synchronization overhead [12].”
Wahed, M. K. Y. A. (2025) · Journal of Information & Knowledge Management · doi:10.1142/s021964922550073x · Cited by 7
Federated Quantum Machine Learning for Distributed Cybersecurity in Multi-Agent Energy Systems
Addo, K., Kabeya, M., & Ojo, E. E. (2025) · Energies · doi:10.3390/en18205418 · Cited by 13
Adaptive Power Control and Data Sampling for Energy-Efficient Over-the-Air Federated Edge Learning
Wang, X., Chen, Y., Wang, B., Ning, Z., Guo, L., Niyato, D. T., & Zhang, Y. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3621448 · Cited by 8
“Despite its potential in mobile edge environments, FL faces significant communication bottlenecks due to the growing scale of model parameters and the participation of numerous distributed devices, which strain limited wireless bandwidth and resources [5], [6].”
Next-generation drug discovery: The AI revolution in pharmaceutical research
Bilotta, M., Rocca, R., & Alcaro, S. (2025) · Artificial Intelligence in the Life Sciences · doi:10.1016/j.ailsci.2025.100149 · Cited by 17
The Impact of Federated Learning on Urban Computing
Souza, J. R. F., Oliveira, S. Z. L. N., & Oliveira, H. M. N. D. S. (2024) · Journal of Internet Services and Applications · doi:10.5753/jisa.2024.4006 · Cited by 9
“FL is an innovative machine learning approach that enables model training across a decentralized network of devices or servers while preserving the privacy of individual data on each device [Beltrán et al ., 2023].”
BackgroundFedHFP: A Federated Deep Learning Framework for Heart Failure Prediction
Lohachab, A., & Kumar, K. (2024) · IETE Journal of Research · doi:10.1080/03772063.2024.2428741 · Cited by 12
“Additionally, this underlying decentralized approach effectively minimizes communication overhead by distributing model training across various devices [7], enabling collaboration among healthcare institutions and safeguarding patient confidentiality.”
A 3D Model-Based Framework for Real-Time Emergency Evacuation Using GIS and IoT Devices
Tyagi, N., Singh, J., Singh, S., & Sehra, S. S. (2024) · ISPRS International Journal of Geo-Information · doi:10.3390/ijgi13120445 · Cited by 9
OCD-FL: A Novel Communication-Efficient Peer Selection-Based Decentralized Federated Learning
Masmoudi, N., & Jaafar, W. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3518836 · Cited by 9
“Index Terms —Decentralized federated learning, peer selection, knowledge gain, energy efficiency.”
An energy-efficient decentralized federated learning framework for mobile-IoT networks
Javan, N. T., Gharyeali, E. Z., & Mostafavi, S. (2025) · Computer Networks · doi:10.1016/j.comnet.2025.111233 · Cited by 13
The Digital Transformation of Education: Opportunities and Challenges for Communication
Kayyali, M. (2025) · IGI Global eBooks · doi:10.4018/979-8-3693-9067-2.ch014 · Cited by 8
Ahmed, R. H., Sultana, J., Zahid, S., Habib, M. A., Rauf, A., & Hussain, M. (2025) · IEEE Access · doi:10.1109/access.2025.3622511 · Cited by 7
“As noted by [5], federated learning approaches enable decentralised training of AI models, preserving data privacy while reducing the communication overhead typically associated with traditional centralised training methods.”
Soltani, B., Haghighi, V., Zhou, Y., Sheng, Q. Z., & Yao, L. (2024) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3627673.3679853 · Cited by 14
Industrial Rare Cyber Attack Detection Based on Federated Diffusion-Squeeze Graph Modeling
Li, F., Lin, J., Wang, D., & Yang, H. (2025) · IEEE Transactions on Industrial Cyber-Physical Systems · doi:10.1109/ticps.2025.3533461 · Cited by 8
“Compared with centralized approaches, data processing and analysis in FL can be performed locally on nodes close to data sources, minimizing extensive data transmission over networks and improving system response speed [13].”
Personalized Decentralized Federated Learning: A Privacy-Enhanced and Byzantine-Resilient Approach
Zhang, A., Zhao, P., Lu, W., & Zhang, G. (2025) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2025.3545905 · Cited by 14
“Decentralized federated learning has shown its great advantages in expanding the feasibility of federated learning by eliminating the reliance on the central server [3], [4].”
DeRelayL: Sustainable Decentralized Relay Learning
Duan, H., Ma, T., Qin, Y., Zeng, R., Cai, W., Leung, V. C. M., & Hu, X. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3558544 · Cited by 7
“Moreover, other researchers also investigate decentralized FL [10], as well as blockchain-enabled FL [11], which mainly studies decentralized model parameter aggregation [12], [13] and decentralization-related topics [14], [15], [16], [17], [18], [19], [20], [21], [22], [23].”
Decentralized federated learning meets Physics-Informed Neural Networks
Alfano, G., Greco, S., Mandaglio, D., Parisi, F., Shahbazian, R., & Trubitsyna, I. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113717 · Cited by 10
MMFed: A Multimodal Federated Learning Framework for Heterogeneous Devices
Wang, G., Zhang, Y., Ying, C., Zhang, Q., Xiong, Z., Wang, J., & Yu, G. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3579858 · Cited by 8
“These approaches aim to remove the central server by enabling direct communication and aggregation among participating clients in a fully distributed manner [33].”
Graph-based Gossiping for Communication Efficiency in Decentralized Federated Learning
Nguyen, H. M., Nguyen, T., Donta, P. K., Pirttikangas, S., & Lovén, L. (2025) · International Conference on Computer Communications and Networks (ICCCN) · doi:10.1109/icccn65249.2025.11133868 · Cited by 6
“Furthermore, these insights can make DFL systems more robust to network variability and disturbances, which is essential given the diverse infrastructure of cross-silo setups [10].”
Scalable Multi-Party Collaborative Data Mining Based on Federated Learning
Wang, M., Kang, T., Dai, L., Yang, H., Du, J., & Liu, C. (2025) · International Conference on Computer Science and Blockchain (CCSB) · doi:10.1109/ccsb66722.2025.11154213 · Cited by 7
“In contrast, real-world data often originates from multiple domains and platforms, exhibiting strong heterogeneity in formats, scales, types, and quality of information [8].”
GreenDFL: A framework for assessing the sustainability of Decentralized Federated Learning systems
Feng, C., Celdrán, A. H., Cheng, X., Bovet, G., & Stiller, B. (2025) · Information and Software Technology · doi:10.1016/j.infsof.2025.107937 · Cited by 8
Self-citationRezaie, H., Taheri, R., Nowroozi, E., Hajizadeh, M., Shiaeles, S. N., & Bauschert, T. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3644477 · Cited by 17
“Clients share updates directly with one another in a peer-to-peer (P2P) manner, using graph-based or consensus-driven protocols for aggregation [30], [31].”
A survey of optimization algorithms for differential privacy in Federated Learning
Shan, F., Liu, Y., Fan, L., Mao, Y., Chen, Z., & Li, S. (2025) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2025.103582 · Cited by 9
Information-Theoretic Decentralized Secure Aggregation With Passive Collusion Resilience
Zhang, X., Li, Z., Li, S., Wan, K., Ng, D. W. K., & Caire, G. (2026) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2026.3678509 · Cited by 22
“Compared to conventional centralized FL, DFL offers several benefits [11].”
Semi-Supervised Decentralized Machine Learning With Device-to-Device Cooperation
Jiang, Z., Xu, Y., Xu, H., Wang, Z., Liu, J., & Qiao, C. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3369188 · Cited by 7
“Additionally, DML schemes mitigate the network bottleneck by spreading communications and workloads more evenly among workers, thereby decreasing the likelihood of network congestion or latency [11].”
BackgroundCSTAR-FL: Stochastic Client Selection for Tree All-Reduce Federated Learning
Xu, Z., Maio, A. D., Samikwa, E., & Braun, T. I. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3507381 · Cited by 8
“WhileguaranteeingtheFLmodelaccuracyunderheteroge-neousdatasettingsiscrucialforreal-worldapplications,several recentstudieshaveinvestigatedtheperformanceofD2D-based communicationstructuresforimplementingFedAvg[28].”
Singh, N., & Adhikari, M. (2024) · Ad Hoc Networks · doi:10.1016/j.adhoc.2024.103728 · Cited by 11
Alamer, A. M. A., & Basudan, S. (2025) · IEEE Access · doi:10.1109/access.2025.3545786 · Cited by 8
“FL is a distributed ML architecture that enables each AV i to locally train its dataset ( dt i ) and transmit only the gradient model to the CC-server instead of raw datasets [10], [11].”
Kang, D. H. E., Kim, D., Song, Y., Lee, D., Kwak, H., & Anthony, B. (2024) · Scientific Reports · doi:10.1038/s41598-024-63393-1 · Cited by 6
Fragkou, E., Chini, E., PAPADOPOULOU, M. K. , Papakostas, D., Katsaros, D., & Dustdar, S. (2024) · IEEE Internet Computing · doi:10.1109/mic.2024.3439068 · Cited by 6
“Similarly, a purely distributed model with only device-to-device communication has been described, 9,11 and other efforts can be found in work by Beltran et al., 3 but these works produce excessive communication since they assume an all-to-all communication pattern.”
Resource-Constrained Decentralized Federated Learning via Personalized Event-Triggering
Zehtabi, S., Hosseinalipour, S., & Brinton, C. G. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3635484 · Cited by 10
“However, access to a central server is not practical in some cases; For example, when conducting online training over dynamic intelligent systems, such as smart vehicles that move across a city, they can rely on vehicle-to-vehicle (V2V) links rather than vehicle-to-edge/server communications [9].”
AIDFL: An Information-Driven Anomaly Detector for Data Poisoning in Decentralized Federated Learning
Chen, X., Feng, C., & Wang, S. (2025) · IEEE Access · doi:10.1109/access.2025.3552168 · Cited by 5
“When confronted with limited resource access, concerns about data privacy, and issues related to national policies and regulations [1], these techniques may face problems such as the leakage of sensitive information and violations of privacy.”
Dong, C., Wu, Q., Li, S., Guo, F., Wang, L., & Zhang, W. (2025) · IEEE Transactions on Dielectrics and Electrical Insulation · doi:10.1109/tdei.2025.3568441 · Cited by 5
“Nevertheless, this fully connected topology frequently results in elevated communication expenses and reduced adaptability, particularly when modifications are necessary for adding or removing links between clients [28].”
Argus: Federated Non-Convex Bilevel Learning Over 6G Space-Air-Ground Integrated Network
Liu, Y., Yang, K., Zhu, Y., Yang, K., & Zhao, H. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3582695 · Cited by 8
“Depending on the different approaches for approximating the hypergradient, the gradient-based bi-level optimization approaches can be further categorized into approximate implicit differentiation (AID) based approaches [29] and iterative differentiation (ITD) based approaches [30].”
Wang, S., Tian, Y., Xiong, J., Ma, J., & Zhang, Y. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3587771 · Cited by 9
“Decentralized federated learning (DFL) with verifiable secure aggregation (SA) offers a viable solution for achieving secure and reliable edge intelligence architectures [8], [9], [10].”
Realistic Urban Traffic Generator Using Decentralized Federated Learning for the SUMO Simulator
Bazán-Guillén, A., Beis-Penedo, C., Cajaraville-Aboy, D., Bautista, P. B., Redondo, R. P. D., Llopis, L. J. D. L. C., Vilas, A. F., Igartua, M. A., & Fernández‐Veiga, M. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3597019 · Cited by 6
“We provide a concise yet fully convergence rate guarantee for the proposed algorithm (collaborative PPO training across DFL scenarios) to justify the attainment of a stationary-point optimal solution.”
Sparse Decentralized Federated Learning
Sha, S., Zhou, S., Kong, L., & Li, G. Y. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3603005 · Cited by 5
“In contrast, decentralized federated learning (DFL [3], [4]), as a new setting of FL, eliminates the need for a central server, allowing nodes to communicate directly with their neighbours based on a predefined network topology and perform local updates.”
Autoencoder-based decentralized federated learning for efficient communication
Mamond, A. W., Kundroo, M., & Kim, T. (2025) · Computer Networks · doi:10.1016/j.comnet.2025.111676 · Cited by 7
UA-PDFL: A personalized approach for decentralized federated learning
Zhu, H., Fan, Y., & Xie, Z. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.131579 · Cited by 8
Initialisation and network effects in decentralised federated learning
Badie-Modiri, A., Boldrini, C., Valerio, L., Kertész, J., & Karsai, M. (2025) · Applied Network Science · doi:10.1007/s41109-025-00737-4 · Cited by 6
“Decentralised federated learning [4] comes as a natural next step in the development of the field of federated learning to save communication costs, improve robustness and privacy [1].”
Toward Communication-Efficient Decentralized Federated Graph Learning Over Non-IID Data
Wang, S., Liu, J., Xu, H., Tang, C., Ma, Q., & Huang, L. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3641696 · Cited by 5
“One of the primary concerns in traditional worker-server-based FGL is the single point of failure and limited scalability [61].”
Ma, X., Fang, F., Liu, T., & Wang, X. (2026) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3650377 · Cited by 17
Gradient purification: defense against data poisoning attack in decentralized federated learning
Li, B., Miao, X., Zhang, Y., & Yin, J. (2026) · Frontiers of Computer Science · doi:10.1007/s11704-025-50240-3 · Cited by 5
Centralized vs Decentralized Federated Learning: A trade-off performance analysis
Medjadji, C., Leduc, G., Kubler, S., & Traon, Y. L. (2024) · International Conference on Future Internet of Things and Cloud (FiCloud) · doi:10.1109/ficloud62933.2024.00019 · Cited by 9
“The recent survey of Beltran et al. [5] has proposed a taxonomy of KPIs, along with categories, to cover the different aspects of FL, as shown in Fig.”
InfluentialNoiseless Privacy-Preserving Decentralized Learning
Biswas, S., Even, M., Kermarrec, A.-M., Massoulié, L., Pires, R. P., Sharma, R., & de Vos, M. (2024) · Proceedings on Privacy Enhancing Technologies · doi:10.56553/popets-2025-0043 · Cited by 18
Li, Q., Zhang, M., Sun, T., Yin, Q., & Shen, L. (2024) · Neural Networks · doi:10.1016/j.neunet.2024.107084 · Cited by 6
Robust decentralized federated learning for heterogeneous and non-ideal networks
Li, B., Gao, W., Xie, J., Li, H., & Gong, M. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111362 · Cited by 8
Accelerating decentralized federated learning via momentum GD with heterogeneous delays
Li, N., Shan, H., Song, M., Zhou, Y., Zhao, Z., Yang, H. H., & Hou, F. (2025) · High-Confidence Computing · doi:10.1016/j.hcc.2025.100310 · Cited by 5
Federated Learning Framework for Privacy-Preserving AI in Healthcare
Ammirata, G., Pezzullo, G., Contino, S., Martino, B. D., & Pirrone, R. (2025) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-3-031-87778-0_31 · Cited by 9
FedGP: Genetic Programming for Evolutionary Aggregation in Federated Learning with Non-IID Data
Pacioni, E., de Vega, F. F., & Calvaresi, D. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-90062-4_26 · Cited by 6
TDML — A Trustworthy Distributed Machine Learning Framework
Wang, Z., Wang, Q. L., Yu, G., & Chen, S. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107951 · Cited by 6
Swarm Learning: A Survey of Concepts, Applications, and Trends
Shammar, E. A., Cui, X., & Al‐qaness, M. A. A. (2026) · ACM Transactions on Privacy and Security · doi:10.1145/3799890 · Cited by 16
“Collaborative Learning: Nodes collaborate to train a shared model, benefiting from shared insights without actual data transfer, which is crucial for maintaining patient confidentiality.”
Natural Language Processing using Federated Learning: A Structured Literature Review
Schumann, G., Awick, J.-P., & Gómez, J. M. (2023) · International Conference on Artificial Intelligence, Blockchain, and Internet of Things (AIBThings) · doi:10.1109/aibthings58340.2023.10292481 · Cited by 7
“After the aggregation step, this lead role is then passed on to another node [6] (applicable in decentralized network topologies, Fig.”
BackgroundInfluentialWilhelmi, F., Afraz, N., Guerra, E., & Dini, P. (2024) · Computer Networks · doi:10.1016/j.comnet.2024.110361 · Cited by 7
Shang, W., Ge, J., Ding, L., Jiang, Z., & Sui, H. (2024) · Future Generation Computer Systems · doi:10.1016/j.future.2024.107526 · Cited by 7
Decentralized Learning Based on Gradient Coding With Compressed Communication
Li, C., & Skoglund, M. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3467262 · Cited by 5
“Edge learning can be implemented in two different ways, i.e., distributed learning (DIL) [7] and decentralized learning (DEL) [8].”
BackgroundDecentralized Federated Learning Framework for Social IoT With Dynamic Network Topology
Liang, X., Tang, J., & Siew, M. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3585542 · Cited by 5
“Since the cloud/edge server has a single point of failure risk, the DFL paradigm has emerged in recent years [25], [26], [38], [39], [40], [41], [42].”
Kim, G., & Kim, Y. (2024) · Sensors · doi:10.3390/s24020535 · Cited by 6
“Beltrán et al. [15] examined the key differences between centralized and decentralized FL, focusing on federation architectures, topologies, communication methods, and security approaches.”
BackgroundDesign Pattern and Challenges of Federated Learning with Applications in Industrial Control System
Batool, H., Xu, J., Rehman, A. U., & Hamam, H. (2024) · Journal on artificial intelligence · doi:10.32604/jai.2024.049912 · Cited by 5
“[47], [69] Transfer learning Considers the presence of malicious parties.”
Voyager: MTD-Based Aggregation Protocol for Mitigating Poisoning Attacks on DFL
Feng, C., Celdrán, A. H., Vuong, M., Bovet, G., & Stiller, B. (2024) · IEEE/IFIP Network Operations and Management Symposium (NOMS) · doi:10.1109/noms59830.2024.10575335 · Cited by 6
“To address these concerns, decentralized FL (DFL) has been introduced, where all nodes are considered equal and can act as both trainers and aggregators simultaneously, improving the fault tolerance of the system [2].”
BackgroundSelf-citationZeybek, M. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/addf66 · Cited by 7
Herath, K., Mahangade, S., & Bagchi, S. (2025) · International Conference on Dependable Systems and Networks workshops · doi:10.1109/dsn-w65791.2025.00079 · Cited by 10
“These challenges have motivated the development of alternative architectures, such as decentralized federated learning (DFL) [3], [7], [10]–[15].”
NEBULA - Decentralized Federated Learning for Heterogeneous Networks
Beltrán, E. T. M., Bovet, G., Pérez, G. M., & Celdrán, A. H. (2025) · ACM SIGCOMM Conference Posters and Demos · doi:10.1145/3744969.3748413 · Cited by 6
Self-citationA Survey on Privacy-Preserving Computing in the Automotive Domain
Yuca, N., Matyunin, N., Arzoglou, E., Anagnostopoulos, N. A., & Katzenbeisser, S. (2025) · ACM Computing Surveys · doi:10.1145/3770580 · Cited by 9
Edge AI-based self-learning technique for mitigating DDoS attacks in WSN
Hussain, S., He, J., Zhu, N., Mughal, F. R., Ahmad, S., Hussain, M. I., & Zardari, Z. A. (2025) · Computer Networks · doi:10.1016/j.comnet.2025.111769 · Cited by 7
Ahmed, I., Ahmad, M., & 전광길 (2025) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2025.068319 · Cited by 7
Houichi, M., Jaïdi, F., & Bouhoula, A. (2025) · International Conference on Cybernetics and Intelligent System (ICORIS) · doi:10.1109/icoris67789.2025.11295993 · Cited by 4
“Meanwhile, federated analytics enables decentralized model training or analysis by sharing model parameters rather than raw data, allowing institutions to collaboratively learn from distributed datasets [10].”
Blockchain-enabled defenses in federated learning: A comprehensive survey
Dib, O. (2026) · Computer Science Review · doi:10.1016/j.cosrev.2026.100932 · Cited by 5
Erfan, F., Bellaïche, M., & Halabi, T. (2026) · Internet of Things · doi:10.1016/j.iot.2026.101927 · Cited by 9
Zhou, S., Chen, Y., Liang, W., Li, K., & Zomaya, A. Y. (2026) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2026.3688316 · Cited by 6
“DFL also improves the system’s fault tolerance, scalability, and flexibility [1], [2].”
Metadata
- Authors
- Martínez Beltrán, E. T., Pérez, M. Q., Sánchez, P. M. S., Bernal, S. L., Bovet, G., Pérez, M. G., Pérez, G. M., & Celdrán, A. H.
- Affiliations
- Department of Information and Communications Engineering, University of Murcia, Murcia, Spain
- Universidad de Murcia (ES)
- Bundesamt für Rüstung armasuisse (CH)
- University of Zurich (CH)
- ORCID
- Venue
- IEEE Communications Surveys & Tutorials
- Published
- 2023
- Volume
- 25
- Issue
- 4
- Pages
- 2983-3013
- ISSN
- 1553-877X (ISSN-L)
- 2373-745X
- OpenAlex
- W4386766438
- Cited by
- 625
- References
- 220
- Citations by year
- 2026: 175 · 2025: 282 · 2024: 145 · 2023: 23
- FWCI
- 58.92
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Communications Surveys & Tutorials (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Zurich Open Repository and Archive (University of Zurich) (submittedVersion)
- arXiv (Cornell University)
- Funders
- European Union (PID2021-122466OB-I00, TED2021-129300B-I00)
- the Swiss Federal Office for Defense Procurement (CyberForce, DEFENDIS)
- Universit?t Z?rich
- Fundaci?n S?neca (21629/FPI/21)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Information Systems › Blockchain Technology Applications and Security
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Keywords
- decentralized federated learning, communication mechanisms, performance indicators, federated learning frameworks, application scenarios, open problems
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.