Model Compression and Hardware Acceleration for Neural Networks: A Comprehensive Survey
The DOI matches its Crossref record; 317 of the 334 references it cites were found in a registry or an index, and 1 are not checked yet.
Abstract
Domain-specific hardware is becoming a promising topic in the backdrop of improvement slow down for general-purpose processors due to the foreseeable end of Moore's Law. Machine learning, especially deep neural networks (DNNs), has become the most dazzling domain witnessing successful applications in a wide spectrum of artificial intelligence (AI) tasks. The incomparable accuracy of DNNs is achieved by paying the cost of hungry memory consumption and high computational complexity, which greatly impedes their deployment in embedded systems. Therefore, the DNN compression concept was naturally proposed and widely used for memory saving and compute acceleration. In the past few years, a tremendous number of compression techniques have sprung up to pursue a satisfactory tradeoff between processing efficiency and application accuracy. Recently, this wave has spread to the design of neural network accelerators for gaining extremely high performance. However, the amount of related works is incredibly huge and the reported approaches are quite divergent. This research chaos motivates us to provide a comprehensive survey on the recent advances toward the goal of efficient compression and execution of DNNs without significantly compromising accuracy, involving both the high-level algorithms and their applications in hardware design. In this article, we review the mainstream compression approaches such as compact model, tensor decomposition, data quantization, and network sparsification…
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
“but there is no review discussing all three categories which in future will be integrated seamlessly to design effective and efficient DNNs in the new edge intelligence era.”
From Bringing ai to edge: from deep learning’s perspective (2022) · cites this paper as [28]
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 · cited 3×
“Specifically, an example of the identity connection used in ResNet [30] [see Fig.”
BackgroundMethodVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“In VGG-Nets [27], the interlayer topology is rather simple, where multiple Conv layers are stacked head-to-tail in sequence.”
BackgroundImageNet 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 · cited 3×
“1078 [20] A.”
BackgroundInfluentialHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 2×
“…we discuss recent advances in RNNs that aim to improve the execution efficiency by designing compact RNNs in the level of unit or network. a) Unit level: LSTM is designed to improve the long-term dependence over vanilla RNNs by alleviating the issue of gradient explosion or vanishing [18], [46].”
BackgroundMethodLeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Densely Connected Convolutional Networks
Huang, G., Liu, Z., Van Der Maaten, L., & Weinberger, K. Q. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.243 · Cited by 48,566
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
(2014) · Cited by 33,419
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
MobileNetV2: Inverted Residuals and Linear Bottlenecks
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L.-C. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00474 · Cited by 28,755 · cited 35×
“The feature propagation gradually improves the abstraction level from local to global, which finally enables the network decision such as image recognition. a) Spatial correlation to enlarge RF: Compared to MLP, the key feature of CNNs is the local connections with shared weights, which results in high spatial correlat…”
BackgroundMethodInfluential
Research works printed but not linked: 2
- lin, 2013, Network in network, arXiv 1312 4400
- han, 2015, Learning both weights and connections for efficient neural network, Proc Adv Neural Inf Process Syst, 1135
Cited by
The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century
Maleki Varnosfaderani, S., & Forouzanfar, M. (2024) · Bioengineering · doi:10.3390/bioengineering11040337 · Cited by 931
“This is where hardware acceleration steps in, acting as a powerful engine that fuels AI-powered diagnostics [111].”
BackgroundBeyond Transmitting Bits: Context, Semantics, and Task-Oriented Communications
Gunduz, D., Qin, Z., Aguerri, I. E., Dhillon, H. S., Yang, Z., Yener, A., Wong, K. K., & Chae, C.-B. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3223408 · Cited by 776
“We refer the reader to [149] for a more comprehensive survey of recent advances in model compression techniques for DNNs.”
Cellular, Wide-Area, and Non-Terrestrial IoT: A Survey on 5G Advances and the Road Toward 6G
Vaezi, M., Azari, A., Khosravirad, S. R., Shirvanimoghaddam, M., Azari, M. M., Chasaki, D., & Popovski, P. (2022) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3151028 · Cited by 631
Efficient Acceleration of Deep Learning Inference on Resource-Constrained Edge Devices: A Review
Shuvo, M. M. H., Islam, S. K., Cheng, J., & Morshed, B. I. (2023) · Proceedings of the IEEE · doi:10.1109/jproc.2022.3226481 · Cited by 436
“There are available reviews of different DL techniques [5], DNNs on edge [2], [16], DL compression techniques for hardware acceleration [10], DL inference on FPGA-based platforms [17], edge intelligence [18], and edge computing [19].”
BackgroundMethodHoefler, T., Alistarh, D., Ben-Nun, T., Dryden, N., & Peste, A. (2021) · ETH Zurich · doi:10.3929/ethz-b-000519202 · Cited by 423
Distributed Artificial Intelligence Empowered by End-Edge-Cloud Computing: A Survey
Duan, S., Wang, D., Ren, J., Lyu, F., Zhang, Y., Wu, H., & Shen, X. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3218527 · Cited by 408
“Due to the redundant design of general-purpose processors [162], [163], it becomes difficult to execute DNN models on GPUs.”
MethodCompute-in-Memory Chips for Deep Learning: Recent Trends and Prospects
Yu, S., Jiang, H., Huang, S., Peng, X., & Lu, A. (2021) · IEEE Circuits and Systems Magazine · doi:10.1109/mcas.2021.3092533 · Cited by 391
A Survey on ChatGPT: AI–Generated Contents, Challenges, and Solutions
Wang, Y., Pan, Y., Yan, M., Su, Z., & Luan, T. H. (2023) · IEEE Open Journal of the Computer Society · doi:10.1109/ojcs.2023.3300321 · Cited by 368
“Additionally, model compression techniques [108] such as pruning, quantization, and knowledge distillation can be utilized to reduce the size and complexity of AIGC models, further reducing energy consumption.”
MethodStructured Pruning for Deep Convolutional Neural Networks: A Survey
He, Y., & Xiao, L. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3334614 · Cited by 363
“[26], [46], [47], [48], [49], [50] X X X X X”
Shaygan, M., Meese, C., Li, W., Zhao, X. (., & Nejad, M. (2022) · Transportation Research Part C: Emerging Technologies · doi:10.1016/j.trc.2022.103921 · Cited by 267
A comprehensive review of model compression techniques in machine learning
Dantas, P. V., Sabino da Silva, W., Cordeiro, L. C., & Carvalho, C. B. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05747-w · Cited by 261
“Recent studies have contributed to the field by proposing novel approaches [69, 193, 194], exploring various compression techniques for DNNs, including compact models, tensor decomposition [131, 195], data quantization [122, 196], and network sparsification [197, 198].”
InfluentialEnd-Edge-Cloud Collaborative Computing for Deep Learning: A Comprehensive Survey
Wang, Y., Yang, C., Lan, S., Zhu, L., & Zhang, Y. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3393230 · Cited by 237
“This approach is widely used in existing pruning algorithms, as detailed in recent review literature [133], [134], [135], [136], However, pruning-after-training approaches are less efficient in terms of learning, as the iterative pruning and retraining (fine-tuning) processes consume considerable…”
Hardware Implementation of Deep Network Accelerators Towards Healthcare and Biomedical Applications
Azghadi, M. R., Lammie, C., Eshraghian, J. K., Payvand, M., Donati, E., Linares-Barranco, B., & Indiveri, G. (2020) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2020.3036081 · Cited by 212
“tribute to increasing throughput of a larger processor [81], [82].”
BackgroundA Comprehensive Survey on Model Quantization for Deep Neural Networks in Image Classification
Rokh, B., Azarpeyvand, A., & Khanteymoori, A. (2023) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3623402 · Cited by 232
“The majority of these surveys have reviewed various approaches to DNN compression and acceleration, allocating only a small section to the topic of quantization [82-95].”
BackgroundCapra, M., Bussolino, B., Marchisio, A., Masera, G., Martina, M., & Shafique, M. (2020) · IEEE Access · doi:10.1109/access.2020.3039858 · Cited by 222
“Deng et al. [335] propose a very complete and extensive work that deals thoroughly with the compression of data taking into account the sparsity and quantization.”
BackgroundBeyond Efficiency: A Systematic Survey of Resource-Efficient Large Language Models
Bai, G., Chai, Z., Chen, L., Wang, S., Lu, J., Zhang, N., Shi, T., Yu, Z., Zhu, M., Zhang, Y., et al. (2026) · ACM Computing Surveys · doi:10.1145/3845797 · Cited by 234
Empowering Edge Intelligence: A Comprehensive Survey on On-Device AI Models
Wang, X., Tang, Z., Guo, J., Meng, T., Wang, C., Wang, T., & Jia, W. (2025) · ACM Computing Surveys · doi:10.1145/3724420 · Cited by 202
“The reliance of these models on large parameter counts and powerful processing capabilities necessitates the development of innovative strategies for model compression, optimization, and adaptation to specific operational environments [42].”
Capra, M., Bussolino, B., Marchisio, A., Shafique, M., Masera, G., & Martina, M. (2020) · Future Internet · doi:10.3390/fi12070113 · Cited by 195
Automatic Modulation Classification: A Deep Architecture Survey
Huynh-The, T., Pham, Q.-V., Nguyen, T.-V., Nguyen, T. T., Ruby, R., Zeng, M., & Kim, D.-S. (2021) · IEEE Access · doi:10.1109/access.2021.3120419 · Cited by 183
“Some compression and acceleration techniques [147] can be applied to speed up the training (or learning) and prediction (or inferencing) stages, which are generally categorized into four groups: network pruning, transfer learning, low-rank factorization, and knowledge distillation [148].”
BackgroundMachine and Deep Learning for Resource Allocation in Multi-Access Edge Computing: A Survey
Djigal, H., Xu, J., Liu, L., & Zhang, Y. (2022) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3199544 · Cited by 166
“More details about these CNN models can be found in [77], [84]. c) Recurrent neural networks (RNNs): RNNs contain intralayer recurrent connections, which make them different from MLP and CNN models [85] (Fig.”
BackgroundFlightLLM: Efficient Large Language Model Inference with a Complete Mapping Flow on FPGAs
Zeng, S., Liu, J., Dai, G., Yang, X., Fu, T., Wang, H., Ma, W., Sun, H., Li, S., Huang, Z., et al. (2024) · ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA) · doi:10.1145/3626202.3637562 · Cited by 232
“Model compression methods [13] ( e.g. , sparsification, quantization, etc.) are commonly applied to address the above issues.”
MethodBekamiri, H., Hain, D. S., & Jurowetzki, R. (2024) · Technological Forecasting and Social Change · doi:10.1016/j.techfore.2024.123536 · Cited by 172
Bringing ai to edge: from deep learning’s perspective
Liu, D., Kong, H., Luo, X., Liu, W., & Subramaniam, R. (2022) · Neurocomputing · doi:10.1016/j.neucom.2021.04.141 · Cited by 171
“Some prior DNN reviews cover one of the three categories, like [25] [26] for model compression, [27] for early works on efficient DNN processing and [28] for NAS 1 but there is no review discussing all three categories which in future will be integrated seamlessly to design effective and efficient…”
BackgroundNetLLM: adapting large language models for networking
Wu, D., Wang, X., Qiao, Y., Wang, Z., Jiang, J., Cui, S., & Wang, F. (2024) · Proceedings of the ACM SIGCOMM 2024 Conference · doi:10.1145/3651890.3672268 · Cited by 167
“With this module, NetLLM facilitates efficient acquisition of networking knowledge without relying on the construction of an external knowledge base.”
BackgroundMethodHardware Approximate Techniques for Deep Neural Network Accelerators: A Survey
Armeniakos, G., Zervakis, G., Soudris, D., & Henkel, J. (2022) · ACM Computing Surveys · doi:10.1145/3527156 · Cited by 167
“[14] 2018 [17] 2020 DNN accelerator architectures [116] 2017 Software and hardware optimization for DNN accelerators [12],[27] 2020 [32] 2021 Quantization techniques [71] 2021 Software-based Pruning and Quantization techniques [99] 2021 NN architectures The ever-increasing demand for efficient DNN…”
BackgroundMethodInfluentialFoundation Model for Advancing Healthcare: Challenges, Opportunities and Future Directions
He, Y., Huang, F., Jiang, X., Nie, Y., Wang, M., Wang, J., & Chen, H. (2024) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2024.3496744 · Cited by 245
“Although some model compression and acceleration techniques on AI models have been studied [273], scaling down such large models and maintaining their general capabilities are still challenging.”
Review of Lightweight Deep Convolutional Neural Networks
Chen, F., Li, S., Han, J., Ren, F., & Yang, Z. (2024) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-023-10032-z · Cited by 164
“[15, 17–19] start from the perspective of deep learning, which encompasses a broad range of technologies and applications to provide specific guidance.”
Enable Deep Learning on Mobile Devices: Methods, Systems, and Applications
Cai, H., Lin, J., Lin, Y.-J., Liu, Z., Tang, H.-T., Wang, H., Zhu, L.-G., & Han, S. (2022) · ACM Trans. Design Autom. Electr. Syst. · doi:10.1145/3486618 · Cited by 160
Self-citationMixed-Signal Computing for Deep Neural Network Inference
Murmann, B. (2021) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2020.3020286 · Cited by 160
“Progress in this field heavily relies on hardware-algorithm co-design [5] and has led to a variety of insightful design principles for custom digital DNN chips [6]–[9].”
BackgroundGeneralized Video Anomaly Event Detection: Systematic Taxonomy and Comparison of Deep Models
Liu, Y., Yang, D., Wang, Y., Liu, J., Liu, J., Boukerche, A., Sun, P., & Song, L. (2024) · ACM Computing Surveys · doi:10.1145/3645101 · Cited by 159
“[29] and knowledge distillation [43] can drive the development of lightweight GVAED models.”
BackgroundLightweight Deep Learning: An Overview
Wang, C.-H., Huang, K.-Y., Yao, Y., Chen, J.-C., Shuai, H.-H., & Cheng, W.-H. (2024) · IEEE Consumer Electronics Magazine · doi:10.1109/mce.2022.3181759 · Cited by 159
Ueda, D., Walston, S. L., Fujita, S., Fushimi, Y., Tsuboyama, T., Kamagata, K., Yamada, A., Yanagawa, M., Ito, R., Fujima, N., et al. (2024) · Diagnostic and Interventional Imaging · doi:10.1016/j.diii.2024.06.002 · Cited by 173
Hardware-Aware Neural Architecture Search: Survey and Taxonomy
Benmeziane, H., Maghraoui, K. E., Ouarnoughi, H., Niar, S., Wistuba, M., & Wang, N. (2021) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2021/592 · Cited by 139
“To reduce the memory footprint in edges devices, model compression techniques [Deng et al. , 2020] are usually applied.”
MethodA Comprehensive Survey on Hardware-Aware Neural Architecture SearchPreprint
Benmeziane, H., Maghraoui, K. E., Ouarnoughi, H., Niar, S., Wistuba, M., & Wang, N. (2021) · arXiv.org · doi:10.48550/arxiv.2101.09336 · Cited by 161
“Relevant surveys [14], [27] on model compression classify the optimizations into these four classes: • Compact Model: This technique modifies the standard operations used in DNNs.”
BackgroundMethodAdaptive Inference through Early-Exit Networks: Design, Challenges and Directions
Laskaridis, S., Kouris, A., & Lane, N. D. (2021) · International Workshop on Embedded and Mobile Deep Learning (EMDL) · doi:10.1145/3469116.3470012 · Cited by 203
“Approaches in the literature exploit various approximation and compression methods [71] to reduce the footprint of these models, including quantisation of network weights and activations [16, 46, 59] or weight sparsification and pruning [9, 19, 51].”
MethodEr, M. J., Chen, J., Zhang, Y., & Gao, W. (2023) · Sensors · doi:10.3390/s23041990 · Cited by 151
“Model compression has even influenced the design of neural network accelerators, resulting in the achievement of extreme performance improvements [119].”
BackgroundNeural Pruning via Growing RegularizationPreprint
Wang, H., Qin, C., Zhang, Y., & Fu, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.09243 · Cited by 174
“As deep neural networks advance in recent years LeCun et al. (2015); Schmidhuber (2015), their remarkable effectiveness comes at a cost of rising storage, memory footprint, computing resources and energy consumption Cheng et al. (2017); Deng et al. (2020).”
BackgroundSurvey on Knowledge Distillation for Large Language Models: Methods, Evaluation, and Application
Yang, C., Zhu, Y. T., Wang, L., Wang, Y., Chen, Q., Gao, C., Yan, B., & Chen, Y. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3699518 · Cited by 226
Compression of Deep Learning Models for Text: A Survey
Gupta, M., & Agrawal, P. (2022) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3487045 · Cited by 175
Advancing 6G: Survey for Explainable AI on Communications and Network Slicing
Sun, H., Liu, Y., Al-Tahmeesschi, A., Nag, A., Soleimanpour-Moghadam, M., Canberk, B., Arslan, H., & Ahmadi, H. (2025) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2025.3534626 · Cited by 151
“Techniques such as model compression, quantization, and the use of specialized hardware accelerators can significantly reduce the energy consumption of XAI methods [229], [230].”
Split Learning in 6G Edge Networks
Lin, Z., Qu, G., Chen, X., & Huang, K. (2024) · IEEE Wireless Communications · doi:10.1109/mwc.014.2300319 · Cited by 199
“There are several popular compression techniques [12]: i) Model quantization reduces the bitwidths of both weights and activations (e.g., from full precisions to 8 bits) to lower the training latency and memory requirements; ii) model pruning directly eliminates the number of parameters of DNN.”
MethodA bird's-eye view of deep learning in bioimage analysis
Meijering, E. (2020) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2020.08.003 · Cited by 126
“…models [258] Data fusion strategies [259] Omics applications [260] Efficient DL Parallelization and distribution [261] Compression and acceleration [262] Biomedical imaging applications [263] Explainable DL Interpretable AI approaches [264] Visual analytics tools [265] Responsible DL On replacing…”
BackgroundMethodA survey on hardware accelerators: Taxonomy, trends, challenges, and perspectives
Peccerillo, B., Mannino, M., Mondelli, A., & Bartolini, S. (2022) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2022.102561 · Cited by 148
Learning Slimming SAR Ship Object Detector Through Network Pruning and Knowledge Distillation
Chen, S., Zhan, R., Wang, W., & Zhang, J. (2020) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2020.3041783 · Cited by 127
“At present, model compression [21] is mainly divided into four types, including pruning, quantization, low-rank factorization, and knowledge distillation (KD).”
Dave, S., Baghdadi, R., Nowatzki, T., Avancha, S., Shrivastava, A., & Li, B. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3098483 · Cited by 140
Spiking Neural Network Integrated Circuits: A Review of Trends and Future Directions
Basu, A., Deng, L., Frenkel, C., & Zhang, X. (2022) · Custom Integrated Circuits Conference · doi:10.1109/cicc53496.2022.9772783 · Cited by 119
Deep neural networks compression: A comparative survey and choice recommendations
Marinò, G. C., Petrini, A., Malchiodi, D., & Frasca, M. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.11.072 · Cited by 117
Light-YOLOv4: An Edge-Device Oriented Target Detection Method for Remote Sensing Images
Ma, X., Ji, K., Xiong, B., Zhang, L., Feng, S., & Kuang, G. (2021) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2021.3120009 · Cited by 118
Wang, H., Shao, W., Sun, C., Yang, K., Cao, D., & Li, J. (2024) · Engineering · doi:10.1016/j.eng.2023.10.011 · Cited by 122
Toward Green AI: A Methodological Survey of the Scientific Literature
Barbierato, E., & Gatti, A. (2024) · IEEE Access · doi:10.1109/access.2024.3360705 · Cited by 115
“It involves methods like weight quantization, pruning, knowledge distillation, and compact architecture design [117].”
MethodToward Open-World Electroencephalogram Decoding Via Deep Learning: A comprehensive survey
Chen, X., Li, C., Liu, A., McKeown, M. J., Qian, R., & Wang, Z. J. (2022) · IEEE Signal Processing Magazine · doi:10.1109/msp.2021.3134629 · Cited by 100
“As a result, reducing the cost of memory and computational complexity in deep models, that is, model compression [84] of deep models without significantly decreasing the model performance for EEG analysis becomes an urgent and promising topic.”
BackgroundHu, C., Zhao, C., Shao, H., Deng, J. Y., & Wang, Y. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3433415 · Cited by 105
High-speed emerging memories for AI hardware accelerators
Lu, A., Lee, J., Kim, T., Karim, M. A. U., Park, R. S., Simka, H. S., & Yu, S. (2024) · Nature Reviews Electrical Engineering · doi:10.1038/s44287-023-00002-9 · Cited by 99
Peng, D., Liu, X., Zhang, Y., Guan, H., Li, Y., & Bruzzone, L. (2024) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2024.104282 · Cited by 98
Intelligent Radio Signal Processing: A Survey
Pham, Q., Nguyen, N. T., Huynh‐The, T., Le, L. B., Lee, K., & Hwang, W. (2021) · IEEE Access · doi:10.1109/access.2021.3087136 · Cited by 99
“Over the last few years, DL has employed a large number of techniques and interested readers are invited to refer to a recent survey [183] for more details.”
BackgroundA Review of Artificial Intelligence in Embedded Systems
Zhang, Z., & Li, J. (2023) · Micromachines · doi:10.3390/mi14050897 · Cited by 113
“…aspects need to be considered: (1) choosing a suitable deployment platform, (2) using hardware accelerators, as mentioned in the second section, (3) applying model compression techniques, as described in the third section [10,11], and (4) enhancing hardware support for AI algorithms and models.”
MethodTrusted AI in Multiagent Systems: An Overview of Privacy and Security for Distributed Learning
Ma, C., Li, J., Wei, K., Liu, B., Ding, M., Yuan, L., Han, Z., & Vincent Poor, H. (2023) · Proceedings of the IEEE · doi:10.1109/jproc.2023.3306773 · Cited by 96
“High accuracy of large neural networks are often achieved by paying the cost of hungry memory consumption and complex computational capability, which greatly impedes the deployment and development in distributed systems [264].”
BackgroundRecent Advances on Neural Network Pruning at Initialization
Wang, H., Qin, C., Bai, Y., Zhang, Y., & Fu, Y. (2022) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2022/786 · Cited by 94
“…outdated survey [ Reed, 1993 ] , recent surveys of pruning alone [ Gale et al. , 2019; Blalock et al. , 2020; Hoefler et al. , 2021 ] or as a sub-topic under the umbrella of model compression and acceleration [ Sze et al. , 2017; Cheng et al. , 2018a; Cheng et al. , 2018b; Deng et al. , 2020].”
BackgroundAdvances and Challenges of Single‐Pixel Imaging Based on Deep Learning
Song, K., Bian, Y., Wang, D., Li, R., Wu, K., Liu, H., Qin, C., Hu, J., & Xiao, L. (2024) · Laser & Photonics Review · doi:10.1002/lpor.202401397 · Cited by 103
Zero-shot Adversarial Quantization
Liu, Y., Zhang, W., & Wang, J. (2021) · 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/cvpr46437.2021.00156 · Cited by 108
“As a consequence, model quantization, converting high-precision parameters to low-precision ones, becomes one of the main paradigms in model compression and acceleration [10].”
MethodAl-Haija, Q. S. A., & Droos, A. (2024) · Expert Systems · doi:10.1111/exsy.13726 · Cited by 118
“Deep learning (Deng et al., 2020) computations can be sped up by specialized hardware, such as GPUs or specialized AI accelerators, which lowers processing time & power consumption • Improves model performance.”
Custom Hardware Architectures for Deep Learning on Portable Devices: A Review
Zaman, K. S., Reaz, M. B. I., Ali, S. H. M., Bakar, A. A. A., & Chowdhury, M. E. H. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3082304 · Cited by 101
“Various NAS architectures have been explored to optimize search space, search method, and performance evaluation [47], [48].”
MethodTestability and Dependability of AI Hardware: Survey, Trends, Challenges, and Perspectives
Su, F., Liu, C., & Stratigopoulos, H. (2023) · IEEE Design and Test · doi:10.1109/mdat.2023.3241116 · Cited by 97
“For recent and thorough surveys on accelerator design for ANNs and SNNs, the readers are referred to [3], [4], [6], [7], [8], [9], and [5], respectively.”
BackgroundA review of AI edge devices and lightweight CNN and LLM deployment
Sun, K., Wang, X., Miao, X., & Zhao, Q. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128791 · Cited by 102
Towards Model Compression for Deep Learning Based Speech Enhancement
Tan, K., & Wang, D. (2021) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2021.3082282 · Cited by 104
“Various model compression techniques have been developed in the deep learning community, which can be broadly categorized into two classes [4].”
BackgroundHamrani, A., Agarwal, A. K., Allouhi, A., & McDaniel, D. (2023) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-023-02171-8 · Cited by 87
“Fortunately, various model compression techniques, such as parameter trimming and sharing, low-rank factorization, and knowledge distillation, are already accessible (Cheng et al., 2018; Deng et al., 2020).”
Deep Learning-Based Sign Language Digits Recognition From Thermal Images With Edge Computing System
Breland, D. S., Skriubakken, S. B., Dayal, A., Jha, A., Yalavarthy, P. K., & Cenkeramaddi, L. R. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3061608 · Cited by 81
Ps and Qs: Quantization-Aware Pruning for Efficient Low Latency Neural Network Inference
Hawks, B., Duarte, J. M., Fraser, N. J., Pappalardo, A., Tran, N. V., & Umuroglu, Y. (2021) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2021.676564 · Cited by 108
“Several approaches have been successfully deployed to compress networks (Cheng et al., 2018; Choudhary et al., 2020; Deng et al., 2020).”
BackgroundIntegrated Photonic Tensor Processing Unit for a Matrix Multiply: A Review
Peserico, N., Shastri, B. J., & Sorger, V. J. (2023) · Journal of Lightwave Technology · doi:10.1109/jlt.2023.3269957 · Cited by 84
“the processing unit to the memory and vice versa [31], [32], [33].”
BackgroundKhan, F. H., Pasha, M. A., & Masud, S. (2021) · Micromachines · doi:10.3390/mi12060665 · Cited by 97
“plementation of general-purpose architecture for AI applications because domain-specific hardware is yet an emerging branch of AI [39].”
BackgroundDNN Surgery: Accelerating DNN Inference on the Edge Through Layer Partitioning
Liang, H., Sang, Q., Hu, C., Cheng, D., Zhou, X., Wang, D., Bao, W., & Wang, Y. (2023) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2023.3258982 · Cited by 87
“There are optimization techniques, such as model compression and model quantization [31], that can reduce the NN model”
MethodNeuroSim Simulator for Compute-in-Memory Hardware Accelerator: Validation and Benchmark
Lu, A., Peng, X., Li, W., Jiang, H., & Yu, S. (2021) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2021.659060 · Cited by 74
“State-of-the-art deep neural network (DNN) – based machine learning algorithms have demonstrated remarkable effectiveness for various arti fi cial intelligence applications such as image processing, speech recognition, and language translation (Deng et al., 2020).”
BackgroundTowards Mixed-Precision Quantization of Neural Networks via Constrained Optimization
Chen, W., Wang, P., & Cheng, J. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00530 · Cited by 104
PDAE: Efficient network intrusion detection in IoT using parallel deep auto-encoders
Basati, A., & Faghih, M. M. (2022) · Information Sciences · doi:10.1016/j.ins.2022.03.065 · Cited by 79
Improving Neural Network Efficiency via Post-training Quantization with Adaptive Floating-Point
Liu, F., Zhao, W., He, Z., Wang, Y., Wang, Z., Dai, C., Liang, X., & Jiang, L. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00523 · Cited by 90
“These small quantization errors are accumulated and consequently result in a significant accuracy drop [7].”
BackgroundRan, S., Yang, X., Liu, M., Zhang, Y., Cheng, C., Zhu, H., & Yuan, Y. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3147328 · Cited by 69
Freire, P. J., Napoli, A., Spinnler, B., Anderson, M., Ron, D. A., Schairer, W., Bex, T., Costa, N. M. S., Turitsyn, S. K., & Prilepsky, J. E. (2023) · Journal of Lightwave Technology · doi:10.1109/jlt.2023.3234327 · Cited by 69
Review of machine learning techniques for optimal power flow
Khaloie, H., Dolányi, M., Toubeau, J., & Vallée, F. (2025) · Applied Energy · doi:10.1016/j.apenergy.2025.125637 · Cited by 87
Digital Versus Analog Artificial Intelligence Accelerators: Advances, trends, and emerging designs
Seo, J.-S., Saikia, J., Meng, J., He, W., Suh, H.-S., Anupreetham, A., Liao, Y., Hasssan, A., & Yeo, I. (2022) · IEEE Solid-State Circuits Magazine · doi:10.1109/mssc.2022.3182935 · Cited by 67
“While AI algorithms have largely improved over recent years to achieve high accuracy with a smaller number of operations and fewer weights, stateof-the-art ImageNet algorithms still require billions of MAC operations per single image inference and storage for millions of weights [1].”
BackgroundCompacting Deep Neural Networks for Internet of Things: Methods and Applications
Zhang, K., Ying, H.-Q., Dai, H., Li, L., Peng, Y.-Y., Guo, K., & Yu, H.-F. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3063497 · Cited by 50
“Furthermore, [7] investigates the impact of hardware on model compression, which nevertheless has not appeared in the previous surveys.”
BackgroundConformal Prediction for Natural Language Processing: A Survey
Campos, M., Farinhas, A., Zerva, C., Figueiredo, M. A. T., & Martins, A. F. T. (2024) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00715 · Cited by 123
“Many of these systems have been shown to be biased (Blodgett et al., 2020).”
Vision-Based Learning for Drones: A Survey
Xiao, J., Zhang, R., Zhang, Y., & Feroskhan, M. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3564184 · Cited by 80
“Techniques like model pruning [206], [207], quantization [208], [209], distillation [210], and the development of specialized hardware accelerators can play a pivotal role in this regard.”
Lee, J., Mukhanov, L., Molahosseini, A. S., Minhas, U. I., Yang, H., del Rincón, J. M., Dichev, K., Hong, C.-H., & Vandierendonck, H. (2023) · ACM Computing Surveys · doi:10.1145/3587095 · Cited by 67
“Cheng et al. [35] conducted a survey on the model compression techniques including pruning, low-rank factorization, compact convolution, and knowledge distillation.”
hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices
Fahim, F., Hawks, B., Herwig, C., Hirschauer, J., Jindariani, S., Tran, N., Carloni, L. P., Di Guglielmo, G., Harris, P., Krupa, J., et al. (2021) · arXiv · doi:10.48550/arxiv.2103.05579 · Cited by 62
Tuli, S., Casale, G., & Jennings, N. R. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3177569 · Cited by 67
“Efficient compression of DNNmodels has been a long studied problem in the literature [43].”
BackgroundJi, Y., & Chen, L. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3213650 · Cited by 69
“Since three types of data show different sensitivity to their accuracy [11], we adopt different quantization strategies accordingly.”
BackgroundMethodA Comprehensive Survey of Machine Learning Applied to Radar Signal ProcessingPreprint
Lang, P., Fu, X., Martorella, M., Dong, J., Qin, R., Meng, X., & Xie, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2009.13702 · Cited by 80
“Besides, some prospective trends have been proposed such as model compression [41], security, network architecture search [594], and capsule neural network [25].”
BackgroundMethodInfluentialFrom Near-Sensor to In-Sensor: A State-of-the-Art Review of Embedded AI Vision Systems
Fabre, W., Haroun, K., Lorrain, V., Lepecq, M., & Sicard, G. (2024) · Sensors · doi:10.3390/s24165446 · Cited by 58
Diana, L., & Dini, P. (2024) · Remote Sensing · doi:10.3390/rs16213957 · Cited by 57
Mixture of experts (MoE): A big data perspective
Gan, W., Ning, Z., Qi, Z., & Yu, P. S. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103664 · Cited by 70
Brain-Inspired Computing: Models and Architectures
Parhi, K. K., & Unnikrishnan, N. K. (2020) · IEEE Open Journal of Circuits and Systems · doi:10.1109/ojcas.2020.3032092 · Cited by 65
“A comprehensive survey of the topic can be found in [121], [127].”
BackgroundMultiuser Co-Inference With Batch Processing Capable Edge Server
Shi, W., Zhou, S., Niu, Z., Jiang, M., & Geng, L. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3192613 · Cited by 61
“compression [10] or DNN hardware accelerators [11].”
Intelligent recognition of defects in high‐speed railway slab track with limited dataset
Cai, X., Tang, X., Pan, S., Wang, Y., Yan, H. X., Ren, Y., Chen, N., & Hou, Y. (2023) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13109 · Cited by 56
“Deep learning model compression algorithms usually involve tensor low-rank decomposition, parameter clipping, training adjustment, network structure optimization, and so forth (Choudhary et al., 2020; Deng et al., 2020).”
Generative AI and DT integrated intelligent process planning: a conceptual framework
Xu, Q., Zhou, G., Zhang, C., Chang, F., Cao, Y., & Zhao, D. (2024) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-024-13861-9 · Cited by 63
Federated and edge learning for large language models
Piccialli, F., Chiaro, D., Qi, P., Bellandi, V., & Damiani, E. (2024) · Information Fusion · doi:10.1016/j.inffus.2024.102840 · Cited by 69
“Nevertheless, reducing model parameters from floating-point representation to lower bit widths, this process may lead to a sacrifice in accuracy [206].”
Sun, B., Wu, Z., Hu, Y., & Li, T. (2022) · Neural Networks · doi:10.1016/j.neunet.2022.03.025 · Cited by 59
A Comprehensive Survey on Training Acceleration for Large Machine Learning Models in IoT
Wang, H., Qu, Z., Zhou, Q., Zhang, H., Luo, B., Xu, W., Guo, S., & Li, R. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3111624 · Cited by 64
SPARK: Scalable and Precision-Aware Acceleration of Neural Networks via Efficient Encoding
Liu, F., Yang, N., Li, H., Wang, Z., Song, Z., Pei, S., & Jiang, L. (2024) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca57654.2024.00082 · Cited by 69
“The interaction between algorithm and hardware is a ubiquitous trend in the neural network compression domain [6].”
BackgroundAPAE: an IoT intrusion detection system using asymmetric parallel auto-encoder
Basati, A., & Faghih, M. M. (2021) · Neural Computing and Applications · doi:10.1007/s00521-021-06011-9 · Cited by 60
“Some methods were recently used to lower the complexity of a deep model, including convolution factorization [15] and network compression [16].”
MethodPrivacy-Preserving Deep Learning Based on Multiparty Secure Computation: A Survey
Zhang, Q., Xin, C., & Wu, H. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3058638 · Cited by 53
“As pointed out in many works [45], [178]–[181], the DL models always contain redundancy.”
BackgroundMethod
Metadata
- Authors
- Deng, B. L., Li, G., Han, S., Shi, L., & Xie, Y.
- Affiliations
- Tsinghua University (CN)
- Massachusetts Institute of Technology (US)
- University of California, Santa Barbara (US)
- Venue
- Proceedings of the IEEE
- Published
- 2020-04
- Volume
- 108
- Issue
- 4
- Pages
- 485-532
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W3012561096
- Cited by
- 996
- References
- 334
- Citations by year
- 2026: 131 · 2025: 171 · 2024: 193 · 2023: 174 · 2022: 152 · 2021: 142 · 2020: 31
- FWCI
- 46.83
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- Proceedings of the IEEE (publishedVersion)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- Funders
- National Science Foundation (1725447)
- Beijing Academy of Artificial Intelligence (BAAI), Tsinghua University Initiative Scientific Research Program
- Institute for Guo Qiang, Tsinghua University
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
- Keywords
- neural network compression, neural network accelerators, deep neural network, data quantization, network sparsification, tensor decomposition, compact modeling, memory efficiency, embedded systems, hardware architecture
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.