Wireless Network Intelligence at the Edge
Open accessThe DOI matches its Crossref record; 205 of the 224 references it cites were found in a registry or an index, and 6 are not checked yet.
Abstract
Fueled by the availability of more data and computing power, recent breakthroughs in cloud-based machine learning (ML) have transformed every aspect of our lives from face recognition and medical diagnosis to natural language processing. However, classical ML exerts severe demands in terms of energy, memory, and computing resources, limiting their adoption for resource-constrained edge devices. The new breed of intelligent devices and high-stake applications (drones, augmented/virtual reality, autonomous systems, and so on) requires a novel paradigm change calling for distributed, low-latency and reliable ML at the wireless network edge (referred to as edge ML). In edge ML, training data are unevenly distributed over a large number of edge nodes, which have access to a tiny fraction of the data. Moreover, training and inference are carried out collectively over wireless links, where edge devices communicate and exchange their learned models (not their private data). In a first of its kind, this article explores the key building blocks of edge ML, different neural network architectural splits and their inherent tradeoffs, as well as theoretical and technical enablers stemming from a wide range of mathematical disciplines. Finally, several case studies pertaining to various high-stake applications are presented to demonstrate the effectiveness of edge ML in unlocking the full potential of 5G and beyond.
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
“Examples of adversarial attacks include data poisoning [69, 319], model update poisoning [44, 67], and model evasion attacks [441, 69, 211].”
From Advances and Open Problems in Federated Learning (2021) · cites this paper as [211]
Experimental — accuracy not yet measured
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
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“To cope with this, the vanilla RNN’s hidden layer can be replaced with a long short-term memory (LSTM) unit [74].”
BackgroundElements of Information Theory
Cover, T. M., & Thomas, J. A. (2001) · Wiley eBooks · doi:10.1002/0471200611 · Cited by 36,808
(2014) · Cited by 33,419
Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation
Cho, K., van Merrienboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1179 · Cited by 25,545 · cited 5×
“GRUs, a gating mechanism in RNNs, are commonly used for modeling speech signals and musics.”
BackgroundInfluentialMultilayer feedforward networks are universal approximators
Hornik, K., Stinchcombe, M., & White, H. (1989) · Neural Networks · doi:10.1016/0893-6080(89)90020-8 · Cited by 21,728 · cited 29×
“Here, n is the sample complexity to satisfy a target generalization error with a target reliability 1 − ε. PAC Framework Applications: (a) PAC-VC Bound: If the hypothesis space H is infinite, e.g., in deep NNs or non-parametric models such as GP, one can derive the GE using the VapnikChervonenkis (VC) dimension VC(H).”
BackgroundInfluentialCommunication-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 4×
“FL has been shown to improve communication efficiency by MSI quantization [30], adjusting the MSI update period [28], and optimizing devices’ scheduling policy [31].”
BackgroundMethodInfluentialDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204
Approximation by superpositions of a sigmoidal function
Cybenko, G. (1989) · Mathematics of Control, Signals, and Systems · doi:10.1007/bf02551274 · Cited by 13,883 · cited 1×
“The proof is provided for shallow NNs [83], [84] and recently for deep NNs [85].”
BackgroundFoundations of statistical natural language processing
Manning, C. D., & Schütze, H. (1999) · Cited by 9,816
Playing Atari with Deep Reinforcement Learning
Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., & Riedmiller, M. (2013) · arXiv · doi:10.48550/arxiv.1312.5602 · Cited by 9,471 · cited 1×
“To illustrate its effectiveness, consider the Atari gaming environment [190] whose output action dimension is only {up, down, left, right}, whereas the input state dimension is the entire pixels per frame.”
Research works printed but not linked: 30
- 2018, Apple iPhone Xs Max
- 2018, Snapdragon 845 Mobile Platform
- 2018, Comparison Chart
- oulu, 2018, 6G Flagship
- 2018, Cloud TPU
- panda, 2018, RDMA-Based Networking Technologies and Middleware for Next-Generation Clusters and Data Centers
- 2018, Architecture Specification
- 2016, Feasibility Study on New Services and Markets Technology Enablers Release-14
- 2017, Study on New Radio Access Technology Physical Layer Aspects Release-14
- 2018, Study on Scenarios and Requirements for Next Generation Access Technologies Release 14
- iwabuchi, 2018, Field experiments on 5G ultra-reliable low-latency communication (URLLC), NTT DoCoMo Tech J, 20, 14
- 2017, Minimum Requirements Related to Technical Performance for IMT-2020 Radio Interface(s)
- 2018, Google Pixel 3 XL
- 10.1109/glocom.2016.7842367 · doi:10.1109/glocom.2016.7842367
- 2016, Recommendations for NGMN KPIs and Requirements for 5G
- briman, 2018, Another Look at Alphago vs Lee Sedol The Power Angle
- goodfellow, 2016, Deep Learning
- 2018, Service Requirements for the 5G System Stage 1 Release-14
- chaudhari, 2017, Entropy-SGD: Biasing gradient descent into wide valleys, Proc ICLR, 1
- vepakomma, 2019, Reducing leakage in distributed deep learning for sensitive health data, Proc ICLR, 1
- nakamoto, 2018, Bitcoin A Peer-to-Peer Electronic Cash System
- 2017, Coverage Analysis of LTE-M Category-M1
- metz, 2018, Machine Learning Invades the Real World on Internet Balloons
- weisstein, 2018, Fractional Chromatic Number
- 2015, IMT vision—Framework and overall objectives of the future development of IMT for 2020 and beyond
- 2017, Study on New Radio Access Technology Radio Access Architecture and Interfaces Release-14
- elgabli, 2019, Communication-efficient decentralized machine learning framework for 5G and beyond, Proc Globecom, 1
- 2018, Intel Skylake X
- 2018, Intel SSD DC P3700 Review
- 2017, Augmented and Virtual Reality The First Wave of 5G Killer Apps
Cited by
Items 201–300 of 702
Communication-Efficient and Federated Multi-Agent Reinforcement Learning
Krouka, M., Elgabli, A., Issaid, C. B., & Bennis, M. (2021) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2021.3130993 · Cited by 48
“Owing to the strict and stringent requirements for 5G and beyond applications such as industry 4.0, network edge intelligence is of paramount importance [1].”
BackgroundInfluentialSelf-citationLSTM-based QoS prediction for 5G-enabled Connected and Automated Mobility applications
Barmpounakis, S., Magoula, L., Koursioumpas, N., Khalili, R., Perdomo, J. M., & Manjunath, R. P. (2021) · IEEE 5G World Forum (5GWF) · doi:10.1109/5gwf52925.2021.00083 · Cited by 48
Federated Spectrum Learning for Reconfigurable Intelligent Surfaces-Aided Wireless Edge Networks
Yang, B. F., Cao, X., Huang, C., Yuen, C., Renzo, M. D., Guan, Y. L., Niyato, D. T., Qian, L., & Debbah, M. A. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3178445 · Cited by 48
“In this case, federated learning (FL) has emerged as a suitable solution to leverage personalized dataset from a large number of mobile users in order to collaboratively train a shared DNN in a decentralized way, while achieving differential data locality [10], [11].”
BackgroundSelf-citationDistributed Intelligence in Wireless Networks
Liu, X., Yu, J., Liu, Y., Gao, Y., Mahmoodi, T., Lambotharan, S., & Tsang, D. H. K. (2023) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2023.3265425 · Cited by 48
“…users need to communicate with the central BS back and forth to exchange model parameters for the learning process in a distributed manner, it is urgently necessary to design the optimal resource management solution to tackle the limited communication bandwidth and transmission power [147].”
BackgroundInfluentialToward Semantic Communication Protocols: A Probabilistic Logic Perspective
Seo, S., Park, J., Ko, S., Choi, J., Bennis, M., & Kim, S. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2023.3288268 · Cited by 48
Self-citationDistributed Machine Learning through Heterogeneous Edge Systems
Hu, H., Wang, D., & Wu, C. (2019) · arXiv · doi:10.48550/arxiv.1911.06949 · Cited by 47
“…(Tao and Li 2018), where each edge system locally trains the dataset it collects, and exchanges model parameter updates (i.e., gradients) with each other through parameter servers (Wang et al. 2018; Kone ˇ cn ` y et al. 2016; Park et al. 2018) (a.k.a., geo-distributed data parallel training).”
BackgroundMethodCoded Federated Computing in Wireless Networks with Straggling Devices and Imperfect CSI
Ha, S., Zhang, J., Simeone, O., & Kang, J. (2019) · 2019 IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit.2019.8849507 · Cited by 47
“These include wireless data centers [2], in which the processors are conventional servers, and federated learning [4], [5], in which computing is carried out collaboratively by mobile devices.”
Federated Learning With Multichannel ALOHA
Choi, J., & Pokhrel, S. R. (2020) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2019.2960243 · Cited by 40
“FEDERATED learning [1]–[3] has been extensively studied as a distributed machine learning approach with data privacy.”
BackgroundMutual Information Driven Federated Learning
Uddin, M. P., Xiang, Y., Lu, X., Yearwood, J. L., & Gao, L. (2020) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2020.3040981 · Cited by 47
“It is obvious that local models with similar performance can increase redundancy in models’ aggregation and also increase communication overhead significantly [2].”
BackgroundInfluentialStatistical Tools and Methodologies for Ultrareliable Low-Latency Communication—A Tutorial
López, O. L. A., Mahmood, N. H., Shehab, M. A., Alves, H., Rosabal, O. M., Marata, L., & Latva‐aho, M. (2023) · Proceedings of the IEEE · doi:10.1109/jproc.2023.3328920 · Cited by 47
“…trust in the algorithms; • ensemble learning [44], [45], which involves combining multiple ML models to improve the accuracy and robustness of the predictions; • transfer learning [32], [46], federated learning (FL) [38], and meta-learning [47], which are specially efficient in low-data scenarios.”
BackgroundMethodD2D-Assisted Federated Learning in Mobile Edge Computing Networks
Zhang, X., Liu, Y., Liu, J., Argyriou, A., & Han, Y. (2021) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc49053.2021.9417459 · Cited by 46
“Since this cooperation is instrumental in MEC systems, Edge Intelligence [3] [4] has attracted significant re search efforts [5].”
BackgroundJanbi, N. F., Mehmood, R., Katib, I. A., Albeshri, A. A., Corchado, J. M., & Yiğitcanlar, T. (2022) · Sensors · doi:10.3390/s22051854 · Cited by 46
Deebak, B. D., & Hwang, S. O. (2023) · Electronics · doi:10.3390/electronics12051250 · Cited by 45
“Mobile [Data] Center has a technological infrastructure to provide comprehensive delivery of data packets with better visualization of information via a dedicated mobile application [28].”
BackgroundDistributed artificial intelligence: Taxonomy, review, framework, and reference architecture
Janbi, N. F., Katib, I. A., & Mehmood, R. (2023) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2023.200231 · Cited by 44
Dynamic Compression Ratio Selection for Edge Inference Systems With Hard Deadlines
Huang, X., & Zhou, S. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.2997128 · Cited by 44
“More recently, emerging edge learning technology [10], [11] focuses on deploying machine learning algorithms at the wireless access network edge.”
BackgroundClient-Based Intelligence for Resource Efficient Vehicular Big Data Transfer in Future 6G Networks
Sliwa, B., Adam, R., & Wietfeld, C. (2021) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2021.3060459 · Cited by 44
A Decade of Video Analytics at Edge: Training, Deployment, Orchestration, and Platforms
Bai, T., Zhao, H., Huang, L., Wang, Z., Kim, D. I., & Nallanathan, A. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3563377 · Cited by 43
“Park et al. [19] and Murshed et al. [26] reviewed research efforts of machine learning for edge computing, while the training and inference of deep learning model at the network edge were focused in [17], [18], [20], [29].”
A New Green Prospective of Non-orthogonal Multiple Access (NOMA) for 5G
Basnayake, V., Jayakody, D. N. K., Sharma, V., Sharma, N., Muthuchidambaranathan, P., & Mabed, H. (2020) · Information · doi:10.3390/info11020089 · Cited by 43
“It is identified as a performance enhancement tool which can achieve intelligence in network operations, optimization, maintenance and management in the emerging mobile networks [159–161].”
BackgroundEtengu, R., Tan, S. C., Lee, C. K., Abbou, F. M., & Chuah, T. C. (2020) · IEEE Access · doi:10.1109/access.2020.3022291 · Cited by 43
“4 is a summary of the available DL techniques [24], [75], [77].”
BackgroundDecentralized and Model-Free Federated Learning: Consensus-Based Distillation in Function Space
Taya, A., Nishio, T., Morikura, M., & Yamamoto, K. (2022) · IEEE Transactions on Signal and Information Processing over Networks · doi:10.1109/tsipn.2022.3205549 · Cited by 43
“Distillation methods are applied in centralized FL [8], [9], [10], [11], [12], [24], [25] because of communication efficiency.”
MethodToward Using Reinforcement Learning for Trigger Selection in Network Slice Mobility
Addad, R. A., Dutra, D. L. C., Taleb, T., & Flinck, H. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3078501 · Cited by 43
Multi-Agent Meta-Reinforcement Learning for Self-Powered and Sustainable Edge Computing Systems
Munir, M. S., Tran, N. H. K., Saad, W., & Hong, C. S. (2021) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2021.3057960 · Cited by 42
Reconfigurable Intelligent Surface for Green Edge Inference in Machine Learning
Hua, S., & Shi, Y. (2019) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps45667.2019.9024398 · Cited by 42
“Mobile edge machine learning (edgeML) offers promising solutions to these challenges and stands out as a potential paradigm for processing inference tasks for mobile devices, by pushing more computing and storage resources to the network edge servers (e.g., base stations) [3], [4].”
BackgroundOptimizing the Learning Performance in Mobile Augmented Reality Systems With CNN
He, Y., Ren, J., Yu, G., & Cai, Y. (2020) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2020.2992329 · Cited by 42
“Therefore, wireless network intelligence can be achieved by the combination of wireless networks and AI, which has been regarded as an important research direction for future wireless networks [3].”
BackgroundFedFog: Network-Aware Optimization of Federated Learning Over Wireless Fog-Cloud Systems
Nguyen, V., Chatzinotas, S., Ottersten, B., & Duong, T. Q. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3167263 · Cited by 42
“Federated learning (FL) is an emerging distributed ML framework that can address many challenges in implementing ML over networks [3], [4].”
Security and QoS issues in blockchain enabled next-generation smart logistic networks: A tutorial
Vaghani, A., Sood, K., & Yu, S. (2022) · Blockchain Research and Applications · doi:10.1016/j.bcra.2022.100082 · Cited by 42
Liu, S., Wen, D., Li, D., Chen, Q., Zhu, G., & Shi, Y. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3440581 · Cited by 41
“To implement these intelligent services, it is desirable to deploy well-trained AI models at the network edge, giving rise to a research area called edge AI inference or edge inference [6], [7], [8].”
MethodCustomized Federated Learning for accelerated edge computing with heterogeneous task targets
Jiang, H., Liu, M., Yang, B., Liu, Q., Li, J., & Guo, X. (2020) · Computer Networks · doi:10.1016/j.comnet.2020.107569 · Cited by 41
Machine Learning in Event-Triggered Control: Recent Advances and Open Issues
Sedghi, L., Ijaz, Z., Noor‐A‐Rahim, M., Witheephanich, K., & Pesch, D. (2022) · IEEE Access · doi:10.1109/access.2022.3191343 · Cited by 41
Random sketch learning for deep neural networks in edge computing
Li, B., Chen, P., Liu, H., Guo, W., Cao, X., Du, J., Zhao, C., & Zhang, J. (2021) · Nature Computational Science · doi:10.1038/s43588-021-00039-6 · Cited by 40
Collaboration in Participant-Centric Federated Learning: A Game-Theoretical Perspective
Huang, G., Chen, X., Ouyang, T., Ma, Q., Chen, L., & Zhang, J. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3194198 · Cited by 40
“WITH the rapid development of Internet of Things (IoT) andmobile networks, massive volumes of user data are being generated and geographically scattered over the edges of networks [1], [2], [3].”
BackgroundByzantine-robust variance-reduced federated learning over distributed non-i.i.d. data
Peng, J., Wu, Z., Ling, Q. H., & Chen, T. (2022) · Information Sciences · doi:10.1016/j.ins.2022.10.120 · Cited by 40
Beyond private 5G networks: applications, architectures, operator models and technological enablers
Maman, M., Strinati, E. C., Dinh, L. N., Haustein, T., Keusgen, W., Wittig, S., Schmieder, M., Barbarossa, S., Merluzzi, M., Costanzo, F., et al. (2021) · EURASIP Journal on Wireless Communications and Networking · doi:10.1186/s13638-021-02067-2 · Cited by 40
“Thus, we now present a strategy for dynamic resource allocation for computation offloading of machine learning tasks at the edge, in the new framework of Edge Machine Learning [46, 47].”
MethodCharacterizing User Association Patterns for Optimizing Small-Cell Edge System Performance
Wu, F., Lyu, F., Wu, H., Ren, J., Zhang, Y., & Shen, X. (2022) · IEEE Network · doi:10.1109/mnet.121.2200089 · Cited by 38
“Lastly, the problem of how to deploy intelligence algorithms on the edge side has been studied in [5, 6] to improve the system performance.”
BackgroundEmbedded Artificial Intelligence: A Comprehensive Literature Review
Huang, X., Wang, H., Qin, S., & Tang, S.-K. (2025) · Electronics · doi:10.3390/electronics14173468 · Cited by 38
“Combined with the high-speed, low-latency characteristics of 5G/6G, EAI and edge computing will support more complex distributed intelligent applications, such as Vehicle-to-Everything (V2X), large-scale real-time IoT analytics, and immersive AR/VR [151].”
Liu, P., Jiang, J., Zhu, G., Cheng, L., Jiang, W., Luo, W., Du, Y., & Wang, Z. (2022) · Frontiers of Information Technology & Electronic Engineering · doi:10.1631/fitee.2100538 · Cited by 38
“This has driven the rapid development of an emerging area called edge intelligence, sitting at the intersection of the two disciplines, namely artificial intelligence (AI) and wireless communications [1], [2].”
BackgroundProxy Experience Replay: Federated Distillation for Distributed Reinforcement Learning
Cha, H., Park, J., Kim, H., Bennis, M., & Kim, S. (2020) · IEEE Intelligent Systems · doi:10.1109/mis.2020.2994942 · Cited by 38
Self-citationAmplify-and-Forward Relaying for Hierarchical Over-the-Air Computation
Wang, F., Xu, J., Lau, V. K. N., & Cui, S. R. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3185074 · Cited by 37
“We have ∪ ]–[4], such as smart agriculture and smart traffic management, in which the statistics of raw data from massive sensors distributed at a wide area need to be swiftly computed for finer data analytics or real-time decision making.”
BackgroundTowards Message Brokers for Generative AI: Survey, Challenges, and Opportunities
Saleh, A., Morabito, R., Dustdar, S., Tarkoma, S., Pirttikangas, S., & Lovén, L. (2025) · ACM Computing Surveys · doi:10.1145/3742891 · Cited by 37
“Ever-growing data volumes & connections overwhelm traditional brokers.”
Adaptive Heterogeneous Client Sampling for Federated Learning Over Wireless Networks
Luo, B., Xiao, W., Wang, S., Huang, J., & Tassiulas, L. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3368473 · Cited by 36
“However, such centralized data collection and training fashion can be quite challenging in mobile edge networks due to the limited wireless communication bandwidth for data transmission and data privacy concerns [3], [4], [5].”
BackgroundStable and compact design of Memristive GoogLeNet Neural Network
Ran, H., Wen, S., Shi, K., & Huang, T. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.01.122 · Cited by 36
Joint Superposition Coding and Training for Federated Learning over Multi-Width Neural Networks
Baek, H., Yun, W. J., Kwak, Y., Jung, S., Ji, M.-Y., Bennis, M., Park, J., & Kim, J. (2021) · IEEE Conference on Computer Communications · doi:10.1109/infocom48880.2022.9796733 · Cited by 33
“To reach its full potential, it is essential to scale up the range of federating devices that are often wirelessly connected while having heterogeneous levels of available energy [3], [4].”
BackgroundSelf-citationHybrid Federated and Centralized Learning
Elbir, A. M., Coleri, S., & Mishra, K. V. (2021) · European Signal Processing Conference (EUSIPCO) · doi:10.23919/eusipco54536.2021.9616120 · Cited by 35
Knowledge-Based Ultra-Low-Latency Semantic Communications for Robotic Edge Intelligence
Zeng, Q., Wang, Z., Zhou, Y., Hai, W., Yang, L., & Huang, K. (2024) · IEEE Transactions on Communications · doi:10.1109/tcomm.2024.3511931 · Cited by 34
“This spread of computing from cloud to edge brings the benefits of stream-lined processing of mobile data, protection of user privacy, relieved network traffic, and provisioning of ultra-low-latency access [3].”
Han, P., Liu, Y., & Guo, L. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3048832 · Cited by 34
“Edge cloud networks that consist of multiple interconnected edge servers are essential for the next generation network that bears modern applications, such as A/VR and IoT, promoting the development of intelligence at the edge [12], [13].”
BackgroundEdge Learning with Timeliness Constraints: Challenges and Solutions
Sun, Y., Shi, W., Huang, X., Zhou, S., & Niu, Z. (2020) · IEEE Communications Magazine · doi:10.1109/mcom.001.2000382 · Cited by 34
“Accordingly, ML models can be trained and implemented at the network edge, namely edge learning [3].”
MethodTemporal-Structure-Assisted Gradient Aggregation for Over-the-Air Federated Edge Learning
Fan, D., Yuan, X., & Zhang, Y. A. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118348 · Cited by 34
“uplink communication overhead of high-dimensional model updating over an unreliable wireless medium turns out to be a critical bottleneck for the implementation of FL at the edge [2], [3].”
BackgroundGoal-Oriented Communication for Edge Learning Based On the Information Bottleneck
Pezone, F., Barbarossa, S., & Lorenzo, P. D. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746468 · Cited by 34
Internet of things‐fog computing‐based framework for smart disaster management
Aljumah, A. A., Kaur, A., Bhatia, M., & Ahanger, T. A. (2020) · Transactions on Emerging Telecommunications Technologies · doi:10.1002/ett.4078 · Cited by 33
“In 2019, Park et al 23 discussed the main building blocks of edge ML, numerous architectural neural networks, and underlying trade-offs as well as theoretical and computational enablers from a wide variety of mathematical disciplines.”
BackgroundHosseinalipour, S., Wang, S., Michelusi, N., Aggarwal, V., Brinton, C. G., Love, D. J., & Chiang, M. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3286987 · Cited by 32
“INTRODUCTION Distributed machine learning (ML) over wireless networks has attracted significant attention recently, for applications ranging from keyboard next word prediction to autonomous driving [1], [2].”
BackgroundDelay-Aware Hierarchical Federated Learning
Lin, F. P.-C., Hosseinalipour, S., Michelusi, N., & Brinton, C. G. (2023) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2023.3329024 · Cited by 32
“This has led to an emerging area of distributed ML over the network edge, which exploits the distributed computing power of IoT devices to realize an intelligent edge/fog [8], [9].”
BackgroundQ-GADMM: Quantized Group ADMM for Communication Efficient Decentralized Machine Learning
Elgabli, A., Park, J., Bedi, A. S., Bennis, M., & Aggarwal, V. (2020) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp40776.2020.9054491 · Cited by 31
“Recently, distributed machine learning (ML) at the network edge has received significant attention [1–3].”
Self-citationGuardians of the Deep Fog: Failure-Resilient DNN Inference from Edge to Cloud
Yousefpour, A., Devic, S., Nguyen, B., Kreidieh, A., Liao, A., Bayen, A. M., & Jue, J. P. (2019) · International Workshop on Challenges in Artificial Intelligence and Machine Learning for Internet of Things (AIChallengeIoT) · doi:10.1145/3363347.3363366 · Cited by 31
“other ML techniques, have been a successful choice for IoT applications in a broad spectrum of domains such as computer vision, speech recognition, medical diagnosis, and natural language processing [17, 21]. Deep learning techniques make use of deep neural networks (DNNs). In certain DNN-empowered IoT applications, t…”
BackgroundLearning Low Resource Consumption CNN through Pruning and Quantization
Qi, Q., Lu, Y., Li, J., Wang, J., Sun, H., & Liao, J. (2021) · IEEE Transactions on Emerging Topics in Computing · doi:10.1109/tetc.2021.3050770 · Cited by 30
Communication Efficient Decentralized Learning Over Bipartite Graphs
Issaid, C. B., Elgabli, A., Park, J., Bennis, M., & Debbah, M. A. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3126859 · Cited by 30
“Machine learning is central to emerging mission-critical applications such as autonomous driving, remote surgery, and the fifth-generation (5G) communication systems and beyond [1]–[3].”
BackgroundSelf-citationMachine Learning-Enabled Data Rate Prediction for 5G NSA Vehicle-to-Cloud Communications
Sliwa, B., Schippers, H., & Wietfeld, C. (2021) · IEEE 5G World Forum (5GWF) · doi:10.1109/5gwf52925.2021.00059 · Cited by 30
Content Popularity Prediction in Fog Radio Access Networks: A Federated Learning Based Approach
Wu, Y., Jiang, Y., Bennis, M., Zheng, F., Gao, X., & You, X. (2020) · ICC 2020 - 2020 IEEE International Conference on Communications (ICC) · doi:10.1109/icc40277.2020.9148697 · Cited by 29
“ing the local models in a distributed manner [11].”
BackgroundSelf-citationDONE: Distributed Approximate Newton-type Method for Federated Edge Learning
Dinh, C. T., Tran, N. H. K., Nguyen, T. D., Bao, W., Balef, A. R., Zhou, B. B., & Zomaya, A. Y. (2022) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2022.3146253 · Cited by 29
“Compared with lo-cal computation cost at each edge worker, the cost of edge communication is considerably higher in terms of speed, delay, and energy consumption [7], [3], [1], [8], [9]; hence, it is often considered as a bottleneck specific to federated edge learning.”
BackgroundLiquid Software-Based Edge Intelligence for Future 6G Networks
Yang, T., Qin, M., Cheng, N., Xu, W., & Zhao, L. Y. (2022) · IEEE Network · doi:10.1109/mnet.011.2000654 · Cited by 29
Sun, Y., Zhou, S., Niu, Z., & Gündüz, D. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9839279 · Cited by 29
Security Enhancement by Identifying Attacks Using Machine Learning for 5G Network
Keserwani, H., Rastogi, H., Kurniullah, A. Z., Janardan, S. K., Raman, R., Rathod, V. M., & Gupta, A. (2022) · International Journal of Communication Networks and Information Security (IJCNIS) · doi:10.17762/ijcnis.v14i2.5494 · Cited by 29
Wu, J., Dai, T., Guan, P., Liu, S., Gou, F., Taherkordi, A., Li, Y., & Li, T. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3367946 · Cited by 29
“Given the limited battery life of edge devices in wireless networks, [36], [37], and [38] primarily emphasize energy consumption, which has become a significant concern in FL. Research findings in [39], [40], and [41] demonstrated that increasing the number of clients can enhance FL convergence.”
BackgroundFederated Learning in Active STARS-Aided Uplink Networks
Yue, X., Guo, X., Mu, X., Zhao, J., Yang, P., Mu, J., & Lu, Z. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3590980 · Cited by 28
“One major challenge is to support massive data transfer and private data transfer services with constrained communication resources to meet the smart service requirements of 6 G networks [5], [6].”
Online Learning for Orchestration of Inference in Multi-user End-edge-cloud Networks
Shahhosseini, S., Seo, D., Kanduri, A., Hu, T., Lim, S.-S., Donyanavard, B., Rahmani, A. M., & Dutt, N. D. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3520129 · Cited by 28
“Reinforcement learning is an efective approach to develop such an understanding and interpret the varying dynamics of such systems [29, 32].”
BackgroundDeep Learning Assisted CSI Estimation for Joint URLLC and eMBB Resource Allocation
Khan, H., Butt, M. M., Samarakoon, S., Sehier, P., & Bennis, M. (2020) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops49005.2020.9145297 · Cited by 28
“On the other hand, [11] studies the fundamental blocks of machine learning i.”
BackgroundSelf-citationRobust Networking: Dynamic Topology Evolution Learning for Internet of Things
Chen, N., Qiu, T., Daneshmand, M., & Wu, D. O. (2021) · ACM Transactions on Sensor Networks · doi:10.1145/3446937 · Cited by 28
Beyond 5G Private Networks: the 5G CONNI Perspective
Strinati, E. C., Haustein, T., Maman, M., Keusgen, W., Wittig, S., Schmieder, M., Barbarossa, S., Merluzzi, M., Klessig, H., Giust, F., et al. (2020) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps50303.2020.9367460 · Cited by 27
“One of the goals for the smart factory is to enable machine learning/AI at the edge [5].”
BackgroundFederated Learning via Disentangled Information Bottleneck
Uddin, M. P., Xiang, Y., Lu, X., Yearwood, J. L., & Gao, L. (2022) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2022.3187962 · Cited by 27
“It is observed that transferring clients’ models with similar performance (whether hyperactive or primeval) creates the excessive redundancy in models’ aggregation and raises the communication costs greatly [46].”
BackgroundPinching-Antenna Systems (PASS) Aided Over-the-Air Computation
Lyu, Z., Li, H., Gao, Y., Xiao, M., & Poor, H. V. (2025) · IEEE Communications Letters · doi:10.1109/lcomm.2025.3601439 · Cited by 26
“T HE proliferation of intelligent Internet-of-Things (IoT) devices and edge artificial intelligence (AI) applications is accelerating the evolution of the sixth-generation (6G) wireless networks towards task-oriented information processing for ultra-low-latency services [1], [2], [3].”
Ren, Z., Mukherjee, M., Bennis, M., & Lloret, J. (2020) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2020.3041396 · Cited by 26
“At the same time, machine learning over wireless networks [3]–[5] exploits data analytics to improve network optimization.”
BackgroundSelf-citationFederated Learning-Based Content Popularity Prediction in Fog Radio Access Networks
Jiang, Y., Wu, Y., Zheng, F., Bennis, M., & You, X. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3124586 · Cited by 26
“(4) Federated learning based model integration: If the local popularity prediction models in F-APs have been obtained, a global model consisting of these F-APs can be generated by integrating the local models in a distributed manner [25].”
BackgroundSelf-citation5G-Enabled Distributed Intelligence Based on O-RAN for Distributed IoT Systems
Firouzi, R., & Rahmani, R. (2022) · Sensors · doi:10.3390/s23010133 · Cited by 26
“extreme mobile broadband (eMBB), and massive machine-based communication (mMTC) [22].”
BackgroundWireless Federated Langevin Monte Carlo: Repurposing Channel Noise for Bayesian Sampling and Privacy
Liu, D., & Simeone, O. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3215663 · Cited by 26
Communication-Efficient Online Federated Learning Strategies for Kernel Regression
Gogineni, V. C., Werner, S., Huang, Y.-F., & Kuh, A. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3218484 · Cited by 26
“The above FL approaches assume fixed data batches locally stored at clients, which might be too restrictive in latency-sensitive or dynamic environments [27], [28].”
BackgroundCommunication-Efficient and Distributed Learning Over Wireless Networks: Principles and Applications
Park, J., Samarakoon, S., Elgabli, A., Kim, J., Bennis, M., Kim, S.-L., & Debbah, M. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3055679 · Cited by 25
“With wireless connectivity, distributed learning frameworks have been overviewed in [13], [28], and [29].”
BackgroundInfluentialSelf-citationL-FGADMM: Layer-Wise Federated Group ADMM for Communication Efficient Decentralized Deep Learning
Elgabli, A., Park, J., Ahmed, S., & Bennis, M. (2020) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc45663.2020.9120758 · Cited by 25
“Interest in data-driven machine learning (ML) is on the rise, but difficulties in securing data still remain [1], [2].”
BackgroundSelf-citationMLOps meets Edge Computing: an Edge Platform with Embedded Intelligence towards 6G Systems
Psaromanolakis, N., Theodorou, V., Laskaratos, D., Kalogeropoulos, I., Vlontzou, M. E., Zarogianni, E., & Samaras, G. (2023) · Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit) · doi:10.1109/eucnc/6gsummit58263.2023.10188244 · Cited by 25
“formance of various ML methods at the edge [8], or from the scope of identifying levels of cloud-edge synergy to efficiently tackle both end-to-end latency and energy consumption on the edge [9].”
BackgroundDecentralized Federated Learning With Adaptive Configuration for Heterogeneous Participants
Liao, Y., Xu, Y., Xu, H., Wang, L., Qian, C., & Qiao, C. (2023) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2023.3335403 · Cited by 25
“With the emergence of EC, federated learning (FL) [7], [8], [9], [10] has been developed to perform distributed model training at the network edge or end devices close to the data source.”
BackgroundMulti-hop Federated Private Data Augmentation with Sample Compression
Jeong, E., Oh, S., Park, J., Kim, H., Bennis, M., & Kim, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.06426 · Cited by 17
“To address the lack of training data samples, it is common to locally oversample the original data, e.g., via rotating and masking for visual data [Park et al., 2018], which may however struggle with locally biased samples.”
MethodSelf-citationJeon, J., Park, S., Choi, M., Kim, J., Kwon, Y., & Cho, S. (2020) · Electronics · doi:10.3390/electronics9091359 · Cited by 24
In-Situ Model Downloading to Realize Versatile Edge AI in 6G Mobile Networks
Huang, K., Hai, W., Liu, Z., & Qi, X. (2023) · IEEE Wireless Communications · doi:10.1109/mwc.010.2200546 · Cited by 24
“6G aims to use AI to empower wide-ranging applications of tactile communications — augmented/virtual reality (AR/VR), remote robotic controls, and autonavigation, to name a few [3].”
BackgroundHiding in the Crowd: Federated Data Augmentation for On-Device Learning
Jeong, E., Oh, S., Park, J., Kim, H., Bennis, M., & Kim, S. (2020) · IEEE Intelligent Systems · doi:10.1109/mis.2020.3028613 · Cited by 24
“O N - DEVICE machine learning (ML) has come to fruition thanks to recent advances in mobile computing power [1], but one key bottleneck has remained unchanged – the lack of training samples.”
BackgroundSelf-citationDynamic-Driven Congestion Control and Segment Rerouting in the 6G-Enabled Data Networks
Deebak, B. D., Al-Turjman, F. M., & Alazab, M. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2020.3023944 · Cited by 24
Joint Transceiver Design for Dual-Functional Full-Duplex Relay Aided Radar-Communication Systems
He, Y., Cai, Y., Yu, G., & Wong, K. (2022) · IEEE Transactions on Communications · doi:10.1109/tcomm.2022.3216334 · Cited by 24
“However, with the rapid development of the artificial intelligence technique, network intelligence has become a new trend of wireless communications [1].”
Adaptive resource optimization for edge inference with goal-oriented communications
Binucci, F., Banelli, P., Lorenzo, P. D., & Barbarossa, S. (2022) · EURASIP Journal on Advances in Signal Processing · doi:10.1186/s13634-022-00958-0 · Cited by 24
“In EI, the mobile devices, also called user equipment (UE), need to perform AI/ML tasks by partially offloading their computations to edge servers (ESs), placed at the edge of the wireless network [4].”
BackgroundAccurate and Fast Federated Learning via IID and Communication-Aware GroupingPreprint
Lee, J., Oh, J., Shin, Y., Lee, J.-G., & Yoon, S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.04857 · Cited by 24
“• Limited Communication Challenge: Because several thousands of devices typically participate in federated learning, the training process is massively distributed, thus causing a huge burden on the backbone (wireless) network (Park et al., 2018).”
BackgroundHe, J., Guo, S., Qiao, D., & Lin, Y. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3145360 · Cited by 23
FedPCC: Parallelism of Communication and Computation for Federated Learning in Wireless Networks
Zhang, H., Tian, H., Dong, M., Ota, K., & Jia, J. (2022) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2022.3170471 · Cited by 23
Data-Importance Aware Radio Resource Allocation: Wireless Communication Helps Machine Learning
Liu, Y., Zeng, Z., Tang, W., & Chen, F. (2020) · IEEE Communications Letters · doi:10.1109/lcomm.2020.2996605 · Cited by 23
Robust Federated Averaging via Outlier Pruning
Uddin, M. P., Xiang, Y., Yearwood, J. L., & Gao, L. (2021) · IEEE Signal Processing Letters · doi:10.1109/lsp.2021.3134893 · Cited by 23
An Efficient Heterogeneous Edge-Cloud Learning Framework for Spectrum Data Compression
Wu, G., Zhou, F., Ding, G., Wu, Q., & Li, X. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3153049 · Cited by 23
“Meanwhile, for spectrum data compressing, the precised spatiotemporal correlation and distribution can be learned by DL models [32], which are beneficial for achieving a higher accuracy and compression rate.”
MethodDifferentially Private CutMix for Split Learning with Vision TransformerPreprint
Oh, S., Park, J., Baek, S., Nam, H., Vepakomma, P., Raskar, R., Bennis, M., & Kim, S. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2210.15986 · Cited by 23
“Motivation: Privacy-Preserving Distributed ML for ViT Edge devices such as phones, cameras, and e-health wearables generate the sheer amount of fresh data [1].”
BackgroundSelf-citationEdge Intelligence Over the Air: Two Faces of Interference in Federated Learning
Chen, Z., Yang, H. H., & Quek, T. Q. S. (2023) · IEEE Communications Magazine · doi:10.1109/mcom.004.2200913 · Cited by 22
“, federated edge learning, has gained significant attention [2], [3].”
BackgroundEfficient Multiuser AI Downloading via Reusable Knowledge Broadcasting
Hai, W., Zeng, Q., & Huang, K. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3373015 · Cited by 22
“The optimization is shown to yield significant system improvements in terms of end-to-end latency, accuracy, throughput, or downstream task performance [12], [13], [14], [15].”
BackgroundLearning Optimal Fronthauling and Decentralized Edge Computation in Fog Radio Access Networks
Lee, H., Kim, J., & Park, S.-H. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3068578 · Cited by 22
“Federated learning (FL) algorithms have been recently studied for handling distributed machine learning problems [17], [21].”
BackgroundLinguaLinked: A Distributed Large Language Model Inference System for Mobile DevicesPreprint
Zhao, J., Song, Y., Liu, S., Harris, I. G., & Jyothi, S. A. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2312.00388 · Cited by 22
“Additionally, centralized processing can result in latency, especially when users are geographically distant from servers or in situations of network congestion [18, 37, 46, 43].”
BackgroundFederated Learning in UAV-Assisted MEC Systems: A Comprehensive Survey
Cheriguene, Y., Jaafar, W., & Yanıkömeroğlu, H. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3608657 · Cited by 21
“In [31], the authors emphasized integrating ML at the wireless network edge, thus facilitating dependable and low-latency communications.”
A comprehensive survey on client selections in federated learning
Gouissem, A., Chkirbene, Z., & Hamila, R. (2024) · CRC Press eBooks · doi:10.1201/9781003496724-41 · Cited by 21
Predictive Control and Communication Co-Design: A Gaussian Process Regression Approach
Girgis, A. M., Park, J., Liu, C., & Bennis, M. (2020) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc48557.2020.9154304 · Cited by 21
Self-citationCoded Computation Across Shared Heterogeneous Workers With Communication Delay
Sun, Y., Zhang, F., Zhao, J., Zhou, S., Niu, Z., & Gündüz, D. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3185905 · Cited by 21
“With the development of edge computing [22] and edge intelligence [23], distributed coded computing is also playing an important role at the edge of the wireless network.”
Background
Metadata
- Authors
- Park, J., Samarakoon, S., Bennis, M., & Debbah, M.
- Affiliations
- University of Oulu (FI)
- Université Paris-Saclay (FR)
- Venue
- Proceedings of the IEEE
- Published
- 2019-11
- Volume
- 107
- Issue
- 11
- Pages
- 2204-2239
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W2904648777
- Cited by
- 92
- References
- 224
- Citations by year
- 2026: 8 · 2025: 2 · 2024: 6 · 2023: 10 · 2022: 13 · 2021: 8 · 2020: 29 · 2019: 16
- FWCI
- 9.03
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- Proceedings of the IEEE (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Figshare (submittedVersion, cc-by)
- arXiv (Cornell University)
- Funders
- Academy of Finland (294128)
- 6Genesis Flagship (318927)
- Kvantum Institute Strategic Project SAFARI
- Academy of Finland through the MISSION Project (319759)
- Artificial Intelligence for Mobile Wireless Systems AIMS project at the University of Oulu
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed Sensor Networks and Detection Algorithms
- Keywords
- edge machine learning, wireless networks, distributed learning, resource-constrained edge devices, 5G, federated learning, low-latency inference, autonomous systems, unmanned aerial vehicles, virtual and augmented reality
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.