Deep Learning With Edge Computing: A Review
The DOI matches its Crossref record; 130 of the 141 references it cites were found in a registry or an index.
Abstract
Deep learning is currently widely used in a variety of applications, including computer vision and natural language processing. End devices, such as smartphones and Internet-of-Things sensors, are generating data that need to be analyzed in real time using deep learning or used to train deep learning models. However, deep learning inference and training require substantial computation resources to run quickly. Edge computing, where a fine mesh of compute nodes are placed close to end devices, is a viable way to meet the high computation and low-latency requirements of deep learning on edge devices and also provides additional benefits in terms of privacy, bandwidth efficiency, and scalability. This paper aims to provide a comprehensive review of the current state of the art at the intersection of deep learning and edge computing. Specifically, it will provide an overview of applications where deep learning is used at the network edge, discuss various approaches for quickly executing deep learning inference across a combination of end devices, edge servers, and the cloud, and describe the methods for training deep learning models across multiple edge devices. It will also discuss open challenges in terms of systems performance, network technologies and management, benchmarks, and privacy. The reader will take away the following concepts from this paper: understanding scenarios where deep learning at the network edge can be useful, understanding common techniques for speeding up deep learning inference and performing distributed training on edge devices, and understanding recent trends and opportunities.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“As mentioned in Section 2.1, the progress of DL is very fast but when it comes to application in Edge or IoT devices then a huge gap is noticeable [67].”
From A survey on deep transfer learning to edge computing for mitigating the COVID-19 pandemic (2020) · cites this paper as [67]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
(2016) · Cited by 31,590
Research works printed but not linked: 30
- Jeans, 2019, Related’s Hudson Yards: Smart City or Surveillance City?
- AT&T Multi-Access Edge Computing
- Blog, Edge Computing at Chick-Fil-A
- Goodfellow, 2016, Deep Learning, 1
- 10.1145/3299710.3211336 · DOI · Paper page
- Tensorflow
- Tensorflowlite
- Caffe2
- Caffe
- Pytorch
- Cuda
- Cudnn
- AWS Deeplens
- Deep Learning for Siri’s Voice: On-Device Deep Mixture Density Networks for Hybrid Unit Selection Synthesis
- Google Duplex: An AI System for Accomplishing Real-World Tasks Over the Phone
- Pagespeed Insights: Improve Server Response Time
- 2017, Hey Siri: An On-Device DNN-Powered Voice Trigger for Apple’s Personal Assistant
- Google’s Pixel Buds are no Match for Professional Interpreters
- Jindal, 2018, Enabling Full Body AR With Mask R-CNN2GO
- 10.1145/3081333.3081360 · DOI · Paper page
- Edge TPU.
- VPU.
- NVIDIA EGX Edge Computing Platform
- Qualcomm Neural Processing SDK for AI
- 10.1109/MNET.2018.1700202 · DOI · Paper page
- Arcore Overview
- Teerapittayanon, 2017, BranchyNet: Fast inference via early exiting from deep neural networks, CoRR, abs/1709.01686, 1 · arXiv
- 10.1145/3195970.3196023 · doi:10.1145/3195970.3196023 · Paper page
- Hsieh, Gaia: Geo-distributed machine learning approaching LAN speeds, 629
- 10.1109/NCA.2017.8171350 · DOI · Paper page
Cited by
Items 301–400 of 1,000
Chun, D., Lee, S., & Kim, H. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3348662 · Cited by 27
“If optimized training is possible on personal devices ( e.g. , mobile/edge devices) using large unlabeled datasets, benefits such as lower annotation cost, lower cloud processing cost, and improved model accuracy for personal applications can be provided [11].”
BackgroundHasan, B. T., & Idrees, A. K. (2023) · Springer eBooks · doi:10.1007/978-3-031-50514-0_1 · Cited by 38
Chirped apodized fiber Bragg gratings inverse design via deep learning
Adibnia, E., Ghadrdan, M., & Mansouri-Birjandi, M. A. (2024) · Optics & Laser Technology · doi:10.1016/j.optlastec.2024.111766 · Cited by 26
Machidon, A. L., Krašovec, A., Pejović, V., & Machidon, O. M. (2025) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2025.3536175 · Cited by 28
“This mismatch between the computational demands of computer vision models and the resource limitations of these hardware devices creates a performance bottleneck, hindering real-time execution of image segmentation or object detection applications [11].”
Collaborative Inference via Ensembles on the Edge
Shlezinger, N., Farhan, E., Morgenstern, H., & Eldar, Y. C. (2021) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp39728.2021.9414740 · Cited by 35
Kilcioglu, E., Mirghasemi, H., Stupia, I., & Vandendorpe, L. (2021) · IEEE Access · doi:10.1109/access.2021.3084689 · Cited by 27
Subedi, P., Hao, J., Kim, I. K., & Ramaswamy, L. M. (2021) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud53861.2021.00016 · Cited by 34
“Moreover, model compression techniques [16], e.”
BackgroundRahman, M. J., Morshed, B. I., Harmon, B. E., & Rahman, M. (2021) · Smart Health · doi:10.1016/j.smhl.2021.100249 · Cited by 31
Manzoor, S., Kim, E., Joo, S.-H., Bae, S.-H., In, G.-G., Joo, K.-J., Choi, J.-H., & Kuc, T. (2022) · IEEE Access · doi:10.1109/access.2022.3190538 · Cited by 28
Winograd Convolution for Deep Neural Networks: Efficient Point Selection
Alam, S. A., Anderson, A., Barabasz, B., & Gregg, D. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3524069 · Cited by 29
“It is also a challenge to use resource constrained devices for such resource intensive applications due to the limited processing, memory and/or battery capacity of many edge devices [5, 6]”
BackgroundLi, X., Xu, Z., Fang, F., Fan, Q., Wang, X., & Leung, V. C. M. (2022) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2022.3177649 · Cited by 28
“Unfortunately, processing ASA tasks in mobile devices greatly consumes the computing resource of the device and reduces the QoE of users [22].”
BackgroundFPG-AI: A Technology-Independent Framework for the Automation of CNN Deployment on FPGAs
Pacini, T., Rapuano, E., & Fanucci, L. (2023) · IEEE Access · doi:10.1109/access.2023.3263392 · Cited by 35
Machine learning at the edge for AI-enabled multiplexed pathogen detection
Ganjalizadeh, V., Meena, G. G., Stott, M. A., Hawkins, A. R., & Schmidt, H. (2023) · Scientific Reports · doi:10.1038/s41598-023-31694-6 · Cited by 24
Ravindran, A. (2023) · IoT · doi:10.3390/iot4040021 · Cited by 42
“Other surveys related to edge computing or video analytics are limited in scope to their respective areas [15–17].”
BackgroundZhu, M., Liu, J., Hu, Z., Liu, J., Jiang, X., & Shi, T. (2024) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.102514 · Cited by 26
Distributed digital twins for health monitoring: resource constrained aero-engine fleet management
Hartwell, A., Montana, F. J., Jacobs, W. R., Kadirkamanathan, V., Ameri, N., & Mills, A. R. (2024) · The Aeronautical Journal · doi:10.1017/aer.2024.23 · Cited by 25
Rahman, M., Roy, P., Frizell, S. S., & Qian, L. (2025) · IEEE Access · doi:10.1109/access.2025.3544107 · Cited by 30
(2025) · doi:10.1038/s41598-025-27818-9 · Cited by 24
A Deep Transfer Learning-based Edge Computing Method for Home Health Monitoring
Sufian, A., You, C., & Dong, M. (2021) · Conference on Information Sciences and Systems · doi:10.1109/ciss50987.2021.9400321 · Cited by 31
“Moreover, the DL-based model usually requires large computing power such as GPU-enabled high resource consumption machine which is a big challenge for EC [34].”
BackgroundMethodPutra, R. V. W., Hanif, M. A., & Shafique, M. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.937782 · Cited by 29
“…is the system that runs Artificial Intelligence (AI) algorithms on resource-and energy-constrained computing devices at the edge of the network, i.e., close to the source of data (Chen and Ran, 2019; Shi et al., 2016; Satyanarayanan, 2017; Yu et al., 2018; Liu et al., 2019; Cao et al., 2020).”
BackgroundCompressing convolutional neural networks with hierarchical Tucker-2 decomposition
Gabor, M., & Zdunek, R. (2022) · Applied Soft Computing · doi:10.1016/j.asoc.2022.109856 · Cited by 31
Edge Computing Solutions for Distributed Machine Learning
Marozzo, F., Orsino, A., Talia, D., & Trunfio, P. (2022) · IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech) · doi:10.1109/dasc/picom/cbdcom/cy55231.2022.9927824 · Cited by 34
“Likewise, [6] reviewed emerging techniques to speed up deep learning models on edge devices.”
BackgroundQuantum Inspired Task Optimization for IoT Edge Fog Computing Environment
Ahanger, T. A., Dahan, F., Tariq, U., & Ullah, I. (2022) · Mathematics · doi:10.3390/math11010156 · Cited by 25
Partial Discharge Detection by Edge Computing
Klein, L., Zmij, P., & Krömer, P. (2023) · IEEE Access · doi:10.1109/access.2023.3268763 · Cited by 28
Decentralized Low-Latency Collaborative Inference via Ensembles on the Edge
Malka, M., Farhan, E., Morgenstern, H., & Shlezinger, N. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3497167 · Cited by 37
“…requires all the users to participate in order to carry out inference, and thus requires constant connectivity, while typically inducing increased latency when the DNN partitions must be applied in a sequential manner, e.g., when each partition is a set of layers in a feed-forward network [3, Sec.”
InfluentialPrivacy-preserving activity and health monitoring on databox
Zhao, Y., Haddadi, H., Skillman, S., Enshaeifar, S., & Barnaghi, P. M. (2020) · doi:10.1145/3378679.3394529 · Cited by 31
“The computational resources on the devices make it possible to run many deep-learning applications, including activity recognition at the edge [6] without releasing raw data to the cloud, meanwhile providing lower latency and network traffic.”
BackgroundSiamese Networks for Few-Shot Learning on Edge Embedded Devices
Lungu, I. A., Aimar, A., Hu, Y., Delbruck, T., & Liu, S.-C. (2020) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2020.3033155 · Cited by 24
“Edge artificial intelligence (AI) hardware accelerators usually target inference networks that have been trained offline on large servers [1].”
BackgroundReal-time traffic sign detection network using DS-DetNet and lite fusion FPN
Ren, K., Huang, L., Fan, C., Han, H., & Deng, H. (2021) · Journal of Real-Time Image Processing · doi:10.1007/s11554-021-01102-1 · Cited by 24
Design Considerations for Energy-efficient Inference on Edge Devices
Hanafy, W. A., Molom-Ochir, T., & Shenoy, R. (2021) · ACM International Conference on Future Energy Systems (e-Energy) · doi:10.1145/3447555.3465326 · Cited by 27
“To address the needs of such applications, new edge hardware accelerators have emerged that make it feasible to perform deep learning inference either on the device itself, or a nearby edge node, at a low energy and resources cost [4].”
BackgroundDistributing DNN training over IoT edge devices based on transfer learning
Tanghatari, E., Kamal, M., Afzali‐Kusha, A., & Pedram, M. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.09.045 · Cited by 27
Goal-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
A Survey on Run-time Power Monitors at the Edge
Zoni, D., Galimberti, A., & Fornaciari, W. (2023) · ACM Computing Surveys · doi:10.1145/3593044 · Cited by 29
“The processing components of devices at the edge are in charge of eiciently performing a large variety of computation- and communication-intensive tasks within tight energy constraints [21], and satisfying the energyand power-eiciency requirements is a key aspect of their design for two main reasons.”
BackgroundNTD: Non-Transferability Enabled Deep Learning Backdoor Detection
Li, Y., Ma, H., Zhang, Z., Gao, Y., Abuadbba, A., Xue, M., Fu, A., Zheng, Y., Al-Sarawi, S. F., & Abbott, D. (2023) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2023.3312973 · Cited by 34
“Nontheless, existing backdoor defenses often rely on heavy computation resources [7], [9], [10] or/and ML expertise [7], [9], [10], [20], [57].”
BackgroundJAVP: Joint-Aware Video Processing with Edge-Cloud Collaboration for DNN Inference
Yang, Z., Ji, W., Guo, Q., & Wang, Z. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3613914 · Cited by 84
“However, for the exponentially growing surveillance video data, it is difficult to satisfy tasks with low delay requirements if all tasks are uploaded to the cloud for processing due to limited bandwidth resources [7].”
BackgroundBaumgart, D. C. (2024) · npj Digital Medicine · doi:10.1038/s41746-024-01005-y · Cited by 33
“While federated learning still requires a central coordinator, swarm learning 31 circumvents it by uniting edge computing 32 and blockchain 33 enabled asynchronous convergence.”
Cheng, T., Masuda, Y., Chen, J., Yu, J., & Hashimoto, M. (2020) · Integration · doi:10.1016/j.vlsi.2020.05.002 · Cited by 27
“In edge computing, the edge servers, typically embedded with GPU processors, are placed near the end devices, and they process and analyze the data independently from the cloud or before sending the data to the cloud [1,2].”
BackgroundSelf-citationPhotoplethysmography based atrial fibrillation detection: a continually growing field
Ding, C., Xiao, R., Wang, W., Holdsworth, E., & Hu, X. (2024) · Physiological Measurement · doi:10.1088/1361-6579/ad37ee · Cited by 31
“This poses significant obstacles for wearable technologies that rely on edge computing, as these calculations can quickly deplete battery life, thereby undermining the feasibility of continuous monitoring (Chen and Ran 2019).”
Deep reinforcement learning in edge networks: Challenges and future directions
Hazra, A., Tummala, V. M. R., Mazumdar, N., Sah, D. K., & Adhikari, M. (2024) · Physical Communication · doi:10.1016/j.phycom.2024.102460 · Cited by 34
Edge Deployment of Deep Networks for Visual Detection: A Review
Zhai, H., Du, J., Ai, Y., & Hu, T. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3502539 · Cited by 35
“TensorFlow Lite offers high inference performance on various hardware platforms [15], including mobile CPUs, GPUs, and dedicated ML accelerators, especially for Google Edge TPUs.”
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 Container-based Design Methodology for Robotic Applications on Kubernetes Edge-Cloud architectures
Lumpp, F., Panato, M., Fummi, F., & Bombieri, N. (2021) · Forum on specification and Design Languages · doi:10.1109/fdl53530.2021.9568376 · Cited by 32
Xie, C., Daghero, F., Chen, Y., Castellano, M., Gandolfi, L., Calimera, A., Macii, E., Poncino, M., & Pagliari, D. J. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3263290 · Cited by 23
“This is beneficial for confidentiality, as it reduces the risk of accidental or malicious leakage of sensitive raw data (e.g., images, audio, video, etc) [1], [6].”
BackgroundEfficient asynchronous federated learning with sparsification and quantization
Jia, J., Liu, J., Zhou, C., Tian, H., Dong, M., & Dou, D. (2024) · Concurrency and Computation Practice and Experience · doi:10.1002/cpe.8002 · Cited by 32
“Finally, Section 6 concludes this article.”
SEE: Scheduling Early Exit for Mobile DNN Inference during Service Outage
Wang, Z., Bao, W., Yuan, D., Ge, L., Tran, N. H. K., & Zomaya, A. Y. (2019) · ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems (MSWiM) · doi:10.1145/3345768.3355917 · Cited by 34
FedViT: Federated continual learning of vision transformer at edge
Zuo, X., Luopan, Y., Han, R., Zhang, Q., Liu, C. H., Wang, G., & Chen, L. Y. (2023) · Future Generation Computer Systems · doi:10.1016/j.future.2023.11.038 · Cited by 27
Qiu, S., Malik, M., Ehsan, H., Wang, W., Wang, J., Cheng, R., Wei, W., & Zaheer, Q. (2025) · Journal of Railway Science and Technology · doi:10.1016/j.jrst.2025.08.003 · Cited by 27
EdgeDRNN: Enabling Low-latency Recurrent Neural Network Edge Inference
Gao, C., Rios-Navarro, A., Chen, X., Delbrück, T., & Liu, S. (2020) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas48895.2020.9074001 · Cited by 27
“In edge computing, computations are done locally on enduser devices to reduce latency and protect privacy [5].”
BackgroundElectric Vehicles Plug-In Duration Forecasting Using Machine Learning for Battery Optimization
Chen, Y., Alamin, K. S. S., Pagliari, D. J., Vinco, S., Macii, E., & Poncino, M. (2020) · Energies · doi:10.3390/en13164208 · Cited by 25
“As an alternative to this approach, we propose a solution based on edge computing, where the ML algorithm is directly executed in the charge station processing hardware [35,36].”
BackgroundMethodUltra-compact binary neural networks for human activity recognition on RISC-V processors
Daghero, F., Xie, C., Pagliari, D. J., Burrello, A., Castellano, M., Gandolfi, L., Calimera, A., Macii, E., & Poncino, M. (2021) · ACM International Conference on Computing Frontiers (CF) · doi:10.1145/3457388.3458656 · Cited by 27
“Moreover, it is also potentially beneficial for energy and privacy, as it avoids the powerhungry transmission of large chunks of (possibly sensitive) user data through wireless channels [7, 28, 31].”
BackgroundAnomalous Vehicle Recognition in Smart Urban Traffic Monitoring as an Edge Service
Chen, N., & Chen, Y. (2022) · Future Internet · doi:10.3390/fi14020054 · Cited by 25
“However, it is non-trivial to fit deep learning algorithms in the edge computing paradigm, mainly because of the limited computational resources at the network edge [35,36].”
BackgroundAdaSTE: An Adaptive Straight-Through Estimator to Train Binary Neural Networks
Le, H., Høier, R. K., Lin, C.-T., & Zach, C. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00055 · Cited by 30
“Deploying deep neural networks (DNNs) to computing hardware such as mobile and IoT devices with limited computational and storage resources is becoming increasingly relevant in practice, and hence training methods especially dedicated to quantized DNNs have emerged as important research topics in recent years [9].”
BackgroundReconstruction of Missing Ground-Penetrating Radar Traces Using Simplified U-Net
Rohman, B. P. A., Nishimoto, M., & Ogata, K. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2021.3072028 · Cited by 21
Binarized Neural Architecture Search for Efficient Object Recognition
Chen, H., Zhuo, L., Zhang, B., Zheng, X., Liu, J., Ji, R., Doermann, D., & Guo, G. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-020-01379-y · Cited by 26
“This makes applications faster and users friendly [9].”
BackgroundDLSTM: Distributed Long Short-Term Memory Neural Networks for the Internet of Things
Wen, G., Qin, J., Fu, X., & Yu, W. (2021) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2021.3054244 · Cited by 21
Bin, K., Lin, J., & Tong, X. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2021.3055795 · Cited by 22
Test Architecture for Systolic Array of Edge-Based AI Accelerator
Solangi, U. S., Ibtesam, M., Ansari, M. A., Kim, J., & Park, S. (2021) · IEEE Access · doi:10.1109/access.2021.3094741 · Cited by 25
“Moreover, there is a growing interest in convergence of DNN processing capabilities with the edge computing devices to enhance application paradigm [4], [5].”
BackgroundBoris, I., Barashok, K., Choi, Y., Choi, Y., Aslam, M., & Lee, J. (2025) · Advances in Mechanical Engineering · doi:10.1177/16878132251347390 · Cited by 24
“Alongside this, edge computing 348–350 is expected to play a growing role in NDE and SHM by enabling real-time processing and inference directly at the point of data acquisition, reducing latency and bandwidth requirements in large-scale deployments.”
Ghaseminya, M. M., Eslami, E., Fazeli, S. A. S., Abouei, J., Abbasi, E., & Karbassi, S. M. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-07799-2 · Cited by 37
Emotion Recognition on Edge Devices: Training and Deployment
Pandelea, V., Ragusa, E., Apicella, T., Gastaldo, P., & Cambria, E. (2021) · Sensors · doi:10.3390/s21134496 · Cited by 23
“The most active research line for training on constrained environment envision an edge server that coordinates the training process with multiple edge devices [34].”
BackgroundSparse Federated Learning With Hierarchical Personalization Models
Liu, X., Wang, Q., Shao, Y., & Li, Y. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3318647 · Cited by 27
“Typically, these samples collected on users’ devices, such as mobile phones, are expected to send to a centralized server with mighty computing power to train a deep model [2].”
BackgroundEdge computing and the internet of things on agricultural green productivity
Shi, H., & Li, Q. (2022) · The Journal of Supercomputing · doi:10.1007/s11227-022-04463-x · Cited by 20
“Against such a dilemma, EC technology is invented [18].”
BackgroundHorizontally Distributed Inference of Deep Neural Networks for AI-Enabled IoT
Rodríguez, I., Campos, C., & Fdez‐Riverola, F. (2023) · Sensors · doi:10.3390/s23041911 · Cited by 28
“…the reader through a comprehensive collection of methods and technologies designed to better leverage edge infrastructures for DNN training [11,12,16,17,21,22,29] but primarily for the execution of such DL models [11,12,16,17,19–27 Regarding inference, an extended trend is particularly…”
InfluentialA Design Flow Based on Docker and Kubernetes for ROS-based Robotic Software Applications
Lumpp, F., Panato, M., Bombieri, N., & Fummi, F. (2023) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3594539 · Cited by 28
“To address all these requirements, software for robotic applications has to be configured to run on heterogeneous computing architectures, by which the different software components have to be properly mapped and orchestrated across heterogeneous hierarchical Edge-Cloud computing platforms [6, 7].”
MethodOn-Device Domain Learning for Keyword Spotting on Low-Power Extreme Edge Embedded Systems
Cioflan, C., Cavigelli, L., Rusci, M., de Prado, M., & Benini, L. (2024) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas59952.2024.10595987 · Cited by 38
“The envisioned adaptation should happen fully on-board to achieve a privacy-by-design system while also minimizing the communication energy costs [4].”
BackgroundWearable Edge AI Applications for Ecological Environments
Silva, M. C., da Silva, J. C. F., Delabrida, S. E., Bianchi, A. G. C., Ribeiro, S. P., Silva, J. S., & Oliveira, R. (2021) · Sensors · doi:10.3390/s21155082 · Cited by 20
“According to Chen and Ran [27], the main challenges regarding the usage of machine and deep learning when considering this aspect are latency, scalability, and privacy.”
BackgroundInfluentialFederated Learning for Privacy-Preserving Speaker Recognition
Woubie, A., & Bäckström, T. (2021) · IEEE Access · doi:10.1109/access.2021.3124029 · Cited by 23
“The data samples are often collected on end-devices such as smartphones and the model is trained using a computationally powerful centralized server [4].”
MethodHu, F., Yang, Q., Yang, J., Shao, J., & Wang, G. (2025) · Environmental Modelling & Software · doi:10.1016/j.envsoft.2025.106524 · Cited by 28
Sivasubramanian, K., Charan, R., Narayanan, S., & Anbarasi, L. J. (2025) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2025.1640549 · Cited by 41
“Table 6 show Resource utilization and power consumption metrics are vital for assessing the feasibility of deploying the CNN-SEEIB model on edge devices (Chen and Ran, 2019).”
Scientific Image Restoration Anywhere
Abeykoon, V., Liu, Z., Kettimuthu, R., Fox, G., & Foster, I. T. (2019) · doi:10.1109/xloop49562.2019.00007 · Cited by 21
“Increasingly, researchers want to run deep neural networks on edge devices [24–26], leading to the need to adapt computationally expensive deep networks for resource-constrained environments.”
BackgroundLatency-Driven Model Placement for Efficient Edge Intelligence Service
Lin, P.-Y., Shi, Z., Xiao, Z., Chen, C., & Li, K. (2021) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2021.3109094 · Cited by 19
Local & Federated Learning at the network edge for efficient predictive analytics
Harth, N., Anagnostopoulos, C. B., Voegel, H.-J., & Kolomvatsos, K. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2022.03.030 · Cited by 25
“The closest related work is from Chen et al. [4] proposing an online asynchronous version of FL.”
An Efficient and Robust Cloud-Based Deep Learning With Knowledge Distillation
Tao, Z., Xia, Q., Chen, S., & Li, Q. (2022) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2022.3160129 · Cited by 22
“In recent years, edge intelligence is expected to push learning computations from the cloud to the edge [14], [15], [16], thus enabling various distributed, low-latency, and reliable, intelligent services.”
BackgroundContactless Palmprint Recognition System: A Survey
Alausa, D. W. S., Adetiba, E., Badejo, J. A., Davidson, I. E. A., Obiyemi, O. O., Buraimoh, E., Abayomi, A., & Oshin, O. I. (2022) · IEEE Access · doi:10.1109/access.2022.3193382 · Cited by 19
Lightweight Network DCR-YOLO for Surface Defect Detection on Printed Circuit Boards
Jiang, Y., Cai, M., & Zhang, D. (2023) · Sensors · doi:10.3390/s23177310 · Cited by 21
Nasir, T., & Malik, M. K. (2023) · Information Processing & Management · doi:10.1016/j.ipm.2023.103544 · Cited by 21
LSTM Network-Based Adaptation Approach for Dynamic Integration in Intelligent End-Edge-Cloud Systems
Yang, X., & Esquivel, J. A. (2024) · Tsinghua Science & Technology · doi:10.26599/tst.2023.9010086 · Cited by 19
Bhoi, S. K., Chakraborty, S., Verbrugge, B., Helsen, S., Robyns, S., Baghdadi, M. E., & Hegazy, O. (2024) · Internet of Things · doi:10.1016/j.iot.2024.101158 · Cited by 39
Lawal, O., Shajihan, S. A. V., Mechitov, K. A., & Spencer, B. F. (2024) · Sensors · doi:10.3390/s24175633 · Cited by 19
“…decision-making, the computation needs to be performed on the WSS. Edge computing paradigms seek to deploy computational resources at or near the location where the data are acquired, thus reducing or eliminating the need for large amounts of data to be collected at a centralized hub [23].”
BackgroundJob scheduling for distributed machine learning in optical WAN
Liu, L., Yu, H., Sun, G., Luo, L., Jin, Q., & Luo, S. (2020) · Future Generation Computer Systems · doi:10.1016/j.future.2020.06.007 · Cited by 20
Distributed Learning Applications in Power Systems: A Review of Methods, Gaps, and Challenges
Gholizadeh, N., & Musílek, P. (2021) · Energies · doi:10.3390/en14123654 · Cited by 26
“Although some of the previous review studies explored the potential of distributed learning algorithms for energy trading [22], machine learning methods for power system security and stability [23], deep learning applications in power systems [24,25], and intersection of deep learning and edge computing [26], no previ…”
Jia, Y., Liu, B., Zhang, X., Dai, F., Khan, A. A., Qi, L., & Dou, W. (2024) · ACM Transactions on Sensor Networks · doi:10.1145/3687478 · Cited by 30
Li, J., Lammers, K., Yin, X., Yin, X., He, L., Sheng, J., Lu, R., & Lu, R. (2025) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2025.109908 · Cited by 25
Hu, T., Shen, P., Zhang, Y., Zhang, J., Li, X., Xia, C., Liu, P., Lü, H., Wu, T., & Han, Z. (2025) · Plant Methods · doi:10.1186/s13007-025-01395-4 · Cited by 18
“In the field of image processing, tools such as OpenCV and ImageJ [1, 20] have become indispensable.”
Li, X., Sun, Y., Lin, J., Li, L., Feng, T., & Yin, S. (2025) · Artificial Intelligence Science and Engineering · doi:10.23919/aise.2005.000007 · Cited by 23
When latency matters: measurements and lessons learned
Iorio, M., Risso, F., & Casetti, C. (2021) · ACM SIGCOMM Computer Communication Review · doi:10.1145/3503954.3503956 · Cited by 29
“machine-learning oriented), hence requiring non-negligible time even if specialized hardware, such as GPUs, is available [5].”
BackgroundTanghatari, E., Kamal, M., Afzali‐Kusha, A., & Pedram, M. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.02.011 · Cited by 26
Cooperative Task Execution for Object Detection in Edge Computing: An Internet of Things Application
Amanatidis, P., Karampatzakis, D., Iosifidis, G., Λάγκας, Θ., & Nikitas, A. (2023) · Applied Sciences · doi:10.3390/app13084982 · Cited by 25
“Jiasi Chen et al. in [21] presented an analytical review that provides an overview of deep learning applications like computer vision, natural language processing, augmented reality, etc. which are used in edge computing.”
BackgroundHybrid Blended Deep Learning Approach for Milk Quality Analysis
Mhapsekar, R. U., O’Shea, N., Davy, S., & Abraham, L. (2024) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2024.3369331 · Cited by 22
“[48].”
Yang, B., Zhou, J., Zhang, S., Xing, Y., Jiang, W., & Xu, L. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3446640 · Cited by 22
IoT Sensor Data Analysis and Fusion Applying Machine Learning and Meta-Heuristic Approaches
Saha, A., Chowdhury, C., Jana, M., & Biswas, S. (2020) · Studies in computational intelligence · doi:10.1007/978-3-030-52067-0_20 · Cited by 23
Edge ai: leveraging the full potential of deep learning
Shuvo, M. M. H. (2022) · Studies in Computational Intelligence · doi:10.1007/978-3-031-14748-7_2 · Cited by 18
Spiking Neural Networks in Intelligent Edge Computing
Zhang, G., Feng, L., Zhou, F., Yang, Z., Zhang, Q., Saleh, A., Donta, P. K., & Dehury, C. K. (2024) · IEEE Consumer Electronics Magazine · doi:10.1109/mce.2024.3506502 · Cited by 18
Towards scalable and efficient Deep-RL in edge computing: A game-based partition approach
Dai, H., Wu, J., Wang, Y., & Xu, C. (2022) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2022.06.006 · Cited by 19
NeuroMessenger: Towards Error Tolerant Distributed Machine Learning Over Edge Networks
Wang, S., & Zhang, X. (2022) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom48880.2022.9796695 · Cited by 20
Seamless Transition From Machine Learning on the Cloud to Industrial Edge Devices With Thinger.io
Bustamante, Á. L., Patricio, M. A., Berlanga, A., & Molina, J. M. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3268771 · Cited by 21
POS: An Operator Scheduling Framework for Multi-model Inference on Edge Intelligent Computing
Zhang, Z., Li, H., Zhao, Y., Lin, C., & Liu, J. (2023) · ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN) · doi:10.1145/3583120.3586953 · Cited by 21
“As we all know, AI accelerators and lightweight deep neural network (DNN) have boosted rapid expansion of edge intelligence applications [1, 10], from intelligent inspection in drones [14] to crop growth monitoring in smart agriculture [39], from voice assistants [36] and recommendation systems [4] in smartphones to v…”
BackgroundElectrodeNet—A Deep-Learning-Based Sound Coding Strategy for Cochlear Implants
Huang, E. H.-H., Chao, R., Tsao, Y., & Wu, C. (2023) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3275587 · Cited by 21
“While current network models appear to require relatively large computational costs, advances in integrated circuits should overcome these limitations [49].”
BackgroundBoosting DNN Cold Inference on Edge Devices
Yi, R., Cao, T., Zhou, A., Ma, X., Wang, S., & Xu, M. (2023) · International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3581791.3596842 · Cited by 27
“, smartphones, IoT, wearables, and autonomous vehicles [22].”
Mukhamediya, A., Fazli, S., & Zollanvari, A. (2023) · IEEE Access · doi:10.1109/access.2023.3287093 · Cited by 27
DistHD: A Learner-Aware Dynamic Encoding Method for Hyperdimensional Classification
Wang, J., Huang, S., & Imani, M. (2023) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac56929.2023.10247876 · Cited by 37
“However, accommodating the high resource requisite of traditional learning methodologies on less powerful computing platforms remains a critical challenge to surmount [1].”
Background
Metadata
- Authors
- Chen, J., & Ran, X.
- Affiliations
- University of California, Riverside (US)
- ORCID
- Chen, Jiasi
- Venue
- Proceedings of the IEEE
- Published
- 2019-08
- Volume
- 107
- Issue
- 8
- Pages
- 1655-1674
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W2960833983
- Cited by
- 1,540
- References
- 141
- Citations by year
- 2026: 205 · 2025: 250 · 2024: 270 · 2023: 245 · 2022: 232 · 2021: 218 · 2020: 114 · 2019: 6
- FWCI
- 77.78
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- Proceedings of the IEEE (publishedVersion)
- Funders
- National Science Foundation (CNS-1817216)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Information Systems › Internet of Things and AI
- Keywords
- edge computing, deep learning, edge inference, distributed deep learning training, edge servers, cloud-edge collaboration, Internet of Things, low latency, privacy, spectral efficiency
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.