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
Items 901–1,000 of 1,002
Edge AI challenges and future directions
Mahalle, P. N., Diwakar, M. P., Ghule, V. R., & Ingle, Y. S. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-26720-8.00005-0
A Novel Edge-Cloud Cooperative System Utilizing Swin-YOLOv11 for Autonomous Vehicle
Alif, M. A., Wang, X., Islam, M. A., Rakib, M. Y. K., Rahman, M. T., & Rahman, M. M. (2025) · IEEE International Conference on Predictive Control of Electrical Drives and Power Electronics (PRECEDE) · doi:10.1109/precede63178.2025.11131088
“Quantization [27] further improves efficiency by reducing the precision of the model’s weights and activations from 32-bit floating-point numbers to lower-precision formats, such as 8-bit integers.”
Learning before filtering: real-time hardware learning at the detector level
Maček, B. (2025) · Journal of Instrumentation · doi:10.1088/1748-0221/20/09/p09014
“8 An effective neuron is defined as the ensemble of all components required to perform both forward and backward propagation for a single neuron.”
Challenges and Opportunities in Federated Learning for the Internet of Everything (IoE)
Bansal, S., Kumar, P., raj, T. P., Mishra, S., & Goyal, P. (2025) · Studies in systems, decision and control · doi:10.1007/978-3-031-99270-4_7 · Cited by 1
A Lightweight CBDNet Hardware Accelerator for Image Denoising
Zhang, R., Li, H., Qiu, Z., Guan, Z., & Wang, Y. (2025) · International Conference on Electronic Communication and Artificial Intelligence (ICECAI) · doi:10.1109/icecai66283.2025.11170485
“To address this challenge, hardware accelerators have gained widespread attention as an effective solution to the high computational complexity and energy consumption of neural networks [9, 10] .”
MDML_KD: Cross-Domain Knowledge Distillation Framework for Bearing Diagnosis
Hung, C., Liao, Z.-J., Wang, C.-C., Chang, T.-A., Chou, Y.-F., & Lee, C. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3618955
“To balance performance and lightweight deployment, model compression techniques have been proposed and widely adopted [11].”
GranQ: Efficient Channel-wise Quantization via Vectorized Pre-Scaling for Zero-Shot QAT
Hong, I., Jo, Y., Lee, H., Ahn, S., Lee, K., & Park, S. (2026) · ACM Symposium on Applied Computing · doi:10.1145/3748522.3779753
“PTQ denotes post-training quantization, and QAT denotes quantization-aware training.”
Model compression using knowledge distillation with integrated gradients
Hernandez, D. E., Chang, J. R., & Nordling, T. E. M. (2026) · Neural Computing and Applications · doi:10.1007/s00521-026-12137-5
“Deploying Deep Learning (DL) models in real-time applications is challenging due to their high computational demands, particularly on edge devices such as smartphones and IoT systems (Szegedy et al., 2017; Deng et al., 2020; Krishnamoorthi, 2018).”
Intelligent Perimeter Intrusion Detection Method Based on Improved YOLOv8 and Edge Computing
Li, X., & He, W. (2025) · International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) · doi:10.1109/icnc-fskd67701.2025.11198065
S-HUB: Scalable Deep Neural Network Fusion for Smart Home Hubs
Yao, Y., Zhao, D. Y., Xu, N., & Zhang, Z. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3626540
“This structured sparsity allows the network to operate directly on general-purpose hardware, eliminating the need for additional hardware design compared to other unstructured pruning methods [35].”
Khurshid, B. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-29981-0.00011-2
ScanNow: A Scan Window-Based Sparse Matrix Multiplication Accelerator Design
Ma, C., Jiang, L. H., Li, Z., Fan, H., Wang, M., Xu, J., & Zhang, W. (2025) · doi:10.1109/iccad66269.2025.11240912
Cross-Entropy Loss of Approximated Deep Neural Networks
Šíma, J., & Vidnerová, P. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-4367-0_31
Block Term Decomposition of Deep Neural Networks by Variational Bayesian
Liu, J., Yang, F., Shi, C., Wu, C., & Liu, W. (2025) · Journal of Shanghai Jiaotong University (Science) · doi:10.1007/s12204-025-2873-x
AI and IoT Applications in Remote Healthcare in the Era of Climate Change: A Review
Ebadinezhad, S. (2025) · IGI Global eBooks · doi:10.4018/979-8-3373-3131-7.ch004
A Hybrid Platform for Multi-Neurons Model with Optimum Co-Design Method
Wang, S., Qiao, K., Liang, J., Hua, N., Lu, R., Zhai, J., & Pei, J. (2025) · Big Data Mining and Analytics · doi:10.26599/bdma.2025.9020106
Cho, S., Kim, H., Kang, D. K., Pyeon, Y. J., Cho, J., Kim, Y., Jung, E. S., Jeong, H. E., & Kim, J. J. (2025) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2025.3642332
“Advances in digital processors have enabled on-device AI systems for hypertension [8], arrhythmia [9], and hazardous gas [10], which have faster feedback, lower power, and enhanced data privacy characteristics [11].”
Hsing, N.-W., Bo-Ya, P., & Chiang, P.-J. (2025) · Applied Optics · doi:10.1364/ao.581780
Dynamic Source-Aware Transfer Learning for Vehicular Platoon Predictive Control
Wu, Y., Zuo, Z., Wang, Y., He, D., Li, J., Zhou, Q., & Xu, H. (2025) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2025.3639826 · Cited by 1
“To reduce computational costs, various model compression techniques such as weight pruning, parameter quantization and model distillation triggered widespread attention [19], [20].”
Nousias, S., Pikoulis, E. V., Mavrokefalidis, C., & Lalos, A. S. (2021) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.4644092
Proseminar Mobile Computing SS 2023 : Mobile und Verteilte Systeme Ubiquitous Computing. Teil XX
Studt, A., Tremper, P., Li, C., Zhao, H., & Beigl, M. (2024) · Repository KITopen (Karlsruhe Institute of Technology) · doi:10.5445/ir/1000170114
Scalable and flexible low-resource service orchestration for enabling smart cities
Goethals, T. (2022) · Ghent University Academic Bibliography (Ghent University)
Liu, K., Fong, S. J., Marques, J. A. L., Lin, W., Millham, R. C., Dey, N., Wang, R., Sciarrone, A., & Qin, H. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3750216
Bayesian policy distillation: Towards lightweight and fast neural policy networks
Kim, J., Jang, Y., Park, J., Gil, Y., & Han, S. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.113539
Self-citationSuminov, K., Kirilyuk, M., Bocharov, N., Paramonov, N., Yanko, D., & Timonov, A. (2025) · International Conference on Control Systems, Mathematical Modeling, Automation and Energy Efficiency (SUMMA) · doi:10.1109/summa68668.2025.11302302
“The quantity of computing modules increases [5], algorithms become more complex, and demands on system adaptability grow [6].”
Evolutionary Algorithm for STDP-Based Spiking Neural Network Model Compression
Yuga, H., Wang, Z., & Dang, K. N. (2025) · IEEE International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC) · doi:10.1109/mcsoc67473.2025.00057
“In recent years, a tremendous number of compression techniques have emerged for Deep Neural Networks [13].”
Adaptive Strategies For Reducing The Carbon Footprint of Deep Neural Networks
Ephrem, T. B., Arrigoni, V., Attenni, G., & Bartolini, N. (2025) · International Conference on Distributed Computing and Networking (ICDCN) · doi:10.1145/3737611.3776967
ВЫБОР ПАРАМЕТРОВ ПРИ СИНТЕЗЕ СИСТЕМЫ УПРАВЛЕНИЯ РТК НА ОСНОВЕ МНОГОВЕРСИОННОГО МОДЕЛИРОВАНИЯ
Суминов, К. А. (2025) · Nanoindustry Russia · doi:10.22184/1993-8578.2025.18.11s.48.54
Precision boundary modeling for area-efficient Block Floating Point accumulation
He, J., Feng, J., Ju, X., Cao, Y., Luo, Z., Yang, J., Yang, J., Li, G., Cheng, J., Chen, D., et al. (2026) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2026.103704
Self-citationMIX-PC: Enabling Efficient DNN Inference with Mixed Numeric Precision Compilation Optimization
Huang, S., Liu, F., Wang, Z., Yang, N., Li, H., Jiang, L., & Guan, H. (2026) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3785473
“However, this approach is hardware-unfriendly and incurs significant overhead [4, 10, 18, 36].”
InfluentialBeacon: Post-Training Quantization With Integrated Grid Selection
Zhang, S., & Saab, R. (2026) · IEEE Signal Processing Letters · doi:10.1109/lsp.2026.3657704
“Quantization [1], [2], [3] has emerged as a mainstream compression technique [4], [5], [6] for deploying LLMs on resource-constrained devices and in other resource-constrained settings.”
A Two-Stage Threat Flow Detection Framework Based on Autoencoder and Support Vector Data Description
Tian, J., Wu, Y., Wu, M., Shen, H., Zhou, L., & Xiong, X. (2025) · International Conference on Wireless Communications and Signal Processing (WCSP) · doi:10.1109/wcsp68525.2025.1010365
“On this basis, the Deep SVDD loss (the second term) then organizes these high-quality representations into a compact, c -centered hypersphere within the latent space [11]–[13].”
Islam, M. S., Jahiduzzaman, F. M., Tanveer, M. S. R., Murshid, M., Mahmoud, A. S., & Salma, S. (2025) · International Conference on Electrical Information and Communication Technology (EICT) · doi:10.1109/eict68394.2025.11355622
“Moreover, the training process for these models can be complex and time-consuming, involving tuning of many parameters, which hinders quick adaptation to new threats [7].”
IRMKD: an application of instance relation matrix in plant disease recognition
Huang, J.-L., Su, J. J., & Cheng, T. (2026) · Frontiers in Bioinformatics · doi:10.3389/fbinf.2026.1761574
“There are also a number of methods to compress network models, including prunin and knowledge distillation (Deng et al., 2020).”
Privacy-Preserving Model Transcription With Differentially Private Synthetic Distillation
Liu, B., Ge, S., Wang, P., Li, S., & Liu, T. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3659110
“An intuitive idea for model re-deployment is model compression [12] whose main objective is to reduce model parameters or storage so that the model can be deployed into resource-limited devices.”
Influence of Parallelism in Vector-Multiplication Units on Correlation Power Analysis
Brosch, M., Probst, M., Kögler, S., & Sigl, G. (2026) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3795524
“Optimization techniques like pruning and quantization are used to reduce the amount of operations and their complexity [10, 15, 16].”
Statically scheduled interconnect for neural network inter-layer accelerating
Zhou, Q., Xu, H., Li, J., Yang, W., & Chen, C. (2026) · Microelectronics Journal · doi:10.1016/j.mejo.2026.107098
NICE: Deep Neural Network Acceleration via Hardware-Friendly Index Assisted Compression
Yang, N., Liu, F., Wang, Z., Li, H., Zhao, H., Liang, X., Jiang, L., & Guan, H. (2026) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3795884
“In the context of sparse accelerators, an intriguing aspect is the amalgamation of compression techniques with hardware design, an approach widely recognized as software-hardware co-design [6, 35, 47].”
QExpander: Pruning-sensitivity-guided adaptive widening for binary neural networks
Gong, H., & Liu, Y. (2026) · Displays · doi:10.1016/j.displa.2026.103397
Conjugational Stability and Symmetry-Invariant Transformations in Artificial Neural Networks
Yotov, K., & Hadzhikolev, E. N. (2026) · Computer Science and Interdisciplinary Research Journal. · doi:10.70862/csir.2026.0301-18
Zhou, C., Ge, L., Tian, X., Liang, J., Zhang, H., Gao, P., Mo, J., & Yin, X. (2026) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2026.3666983
“Systematic model compression approaches are required for deploying deep learning models on resource-limited edge devices [34].”
A Multi Stage Framework for Efficient Deep Neural Network Compression on Edge Devices
Dubey, S. K., & Ahmad, N. (2025) · International Conference on Intelligent Systems and Embedded Design (ISED) · doi:10.1109/ised67359.2025.11405082
Enhancing Edge AI: A Digital Compute-In-Memory SRAM Macro for Vector Matrix Multiplication
Vishwakarma, V., Mishra, S., Gupta, B. B., & Vishvakarma, S. K. (2025) · IEEE International Conference on Emerging Electronics (ICEE) · doi:10.1109/icee67165.2025.11409700
“Deep learning has rapidly expanded into diverse real-world applications such as keyword detection [1], audio recognition [2], image classification [3], and robotics [4].”
Digital Intelligence for Global Good
Saha, S., & Bose, S. (2026) · doi:10.65525/svup.9788199565463.2026.1-140
Optoelectronic extreme learning machine classifier using speckles with engineered statistics
Barboza, A. C. R., de Paula, A., Oriá, M., & Chevrollier, M. (2026) · Optics Communications · doi:10.1016/j.optcom.2026.133127
Method for Preparing Convolutional Neural Networks for Edge Deployment
Prochukhan, D. (2026) · Electronics and Control Systems · doi:10.18372/1990-5548.87.20773
Chen, C., Liu, Y., Xiong, F. B., & Zhang, F. (2025) · IEEE International Conference on Artificial Intelligence, Computer Science and Information Processing (AICSIP) · doi:10.1109/aicsip65423.2025.11427308
Yang, X., Wei, A., & Wang, Z. (2025) · IEEE International Conference of Safe Production and Informatization (IICSPI) · doi:10.1109/iicspi66775.2025.11437951
“Pruning involves removing unnecessary parameters to reduce the size of the model, while quantization involves approximating the value of weights with a reduced number of bits to improve the efficiency of computations [2].”
Kozynets, A., & Демків, Л. І. (2026) · Artificial Intelligence · doi:10.15407/jai2026.01.059
DAF-Net: Dynamic Aggregation and Cycle-Aware Frequency Mechanism for Time Series Forecasting
Dai, L., Yu, J., & Li, X. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-8414-7_24
AdaBit-TasNet: Speech Separation with Inference Adaptable Precision
Elminshawi, M., Chetupalli, S. R., & Habets, E. A. P. (2025) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru65441.2025.11434690
Garbay, T. (2023) · theses.fr (ABES) · doi:10.70675/2b548ac9zfc75z4a72z90dazfc0e0d1f174e · Cited by 2
Apprentissage profond pour le rehaussement de la parole dans les antennes acoustiques ad-hoc
Furnon, N. (2021) · theses.fr (ABES) · doi:10.70675/f911825fz64edz4b55zbba3z409c13ad8e9d
Foundations of networks towards AI
Lesfari, H. (2022) · theses.fr (ABES) · doi:10.70675/a8d7b18ez21eaz4a41zb986zaf44a57c1659
Compression and federated learning : an approach to frugal machine learning
Leconte, L. (2024) · theses.fr (ABES) · doi:10.70675/e20f182fz29f1z4c43z8b62z28efa11eb33d
Embedded-AI-enabled semantic IoT platform for agroecology
Chollet, N. (2023) · theses.fr (ABES) · doi:10.70675/3ff3a4d6z1c42z4ba0za26bzba7fb74ee5ae
Dynamic architectural optimization of artificial neural networks
Maile, K. (2023) · theses.fr (ABES) · doi:10.70675/8a8952e2z422az4072z9397z4629b985f393
Zhao, L., Wang, Y., Du, D., Song, Y., Yu, H., Zhou, W., & Yu, J. (2026) · Optics Communications · doi:10.1016/j.optcom.2026.133222
DMS: Throughput-Aware Distributed Multimodal Model Surgery on Embedded Devices
Cai, Q., Li, K., Zhang, K., Xie, X. N., Tong, X., & Li, K. (2026) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2026.3682748
“First, NN compression [5], [6] reduces the NN’s computational cost, thereby reducing the execution time.”
Pruning, Knowledge Distillation and Quantization of YOLOv8n for Edge-Based Precision Agriculture
Hmache, F., Bellout, A., Zarboubi, M., ba, A. A., Chahba, M., Dliou, A., Saddik, A., & Bouharroud, R. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-18857-1_29
Fed-MET: Memory-Efficient Elastic Training in Federated Learning
Miao, C., Chang, T., Wu, M., Zha, Y., Peng, J., & Wang, X. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463757
Compressed BC-LISTA via Low-Rank Convolutional Decomposition
Wang, H., Kvich, Y., Pérez, E., Roemer, F., & Eldar, Y. C. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462520
Model Deployment for Edge Video Analytics
Bai, T. (2025) · Textbooks in telecommunication engineering · doi:10.1007/978-3-032-14764-6_5
Federated Learning in Edge Networks
Bhrugubanda, M., Chithaluru, P., Narsimhulu, P., Dheeravath, K., & Pasha, A. (2026) · Studies in computational intelligence · doi:10.1007/978-3-032-16711-8_11
Deep Unfolded Parameter Quantization for Multiplier-free MIMO Receivers
Zhang, Y., Xiong, X., Wu, Q., & Liang, Y. (2025) · China Automation Congress (CAC) · doi:10.1109/cac67268.2025.11486753
“However, the parameters learned by these advanced receivers still face considerable storage and computational resource constraints during hardware implementation [9] [10].”
Ramalakshmi, E., Pallerla, J., Goli, V. R., & Gone, V. (2025) · International Conference on Decision Aid Sciences and Applications (DASA) · doi:10.1109/dasa68193.2025.11499117
“Recent advances in quantization [21], [22], parameter-efficient fine-tuning (PEFT) [23], and general model compression [17] enable Large Language Models (LLMs) to run efficiently on modest, local server hardware.”
Energy-Aware Neural Adaptation for Sustainable AI Inference on Resource-Constrained Systems
Patel, T., Parikh, M., Shanbhag, J., Padhy, A. K., & Soni, V. R. (2026) · International Conference on Communications, Information System and Computer Engineering (CISCE) · doi:10.1109/cisce69494.2026.11504582
“Modern state-of-the-art models continue to grow in size and computational requirements [20], [21].”
Kim, J.-G., & Lee, B. (2026) · Applied Sciences · doi:10.3390/app16104896
ASNA: A Novel Adaptive Sparse Neural Architecture for Embedded Systems
Okuwobi, I. P., Liu, J., Raji, O. S., & Abiodun, O. F. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02869-1
“However, these methods often involve trade-offs between accuracy and efficiency, limiting their applicability in scenarios where both high performance and low resource usage are critical (Deng et al., 2020).”
Al-Junaid, A. F. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-01350-w
Diriba, C., & Bonnechère, B. (2026) · Applied Sciences · doi:10.3390/app16104946
FPGA-Based Neuromorphic Architecture for Spiking Neural Network Emulation
Saulquin, A. (2025) · theses.fr (ABES) · doi:10.70675/63c8ec99z8737z480fz9441z84139fe9543c
Ma, J., Zhu, Q., & Dong, Y. (2026) · Aerospace Science and Technology · doi:10.1016/j.ast.2026.112615
Elsheikh, A. A., Fouda, M. E., & Eltawil, A. M. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.134040
Quantum-Inspired Feature Mapping for Edge AI: A MATLAB-Based Hardware-Aware Co-Design Approach
Vimal Rency, A., Kaviarasu, R., & Latha, P. (2026) · International Conference on Artificial Intelligence and Machine Learning Applications Theme: Healthcare and Internet of Things (AIMLA) · doi:10.1109/aimla67915.2026.11522361
A dual-stage approach for explainable time series forecasting in on-field cyber-physical systems
Emmanuele, A., di Torrepadula, F. R., Barbareschi, M., & Mazzocca, N. (2026) · Internet of Things · doi:10.1016/j.iot.2026.101984
Autonomous Control and Artificial Intelligence in Drones: A Comprehensive ReviewPreprint
Alqudsi, Y. S., Alward, A., Koyuncuoglu, Z., Sulaiman, H., & Alyoussef, F. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6810618 · Cited by 1
Liu, Y., Liu, H., Li, Z., Zou, Y., Liu, M., Zong‐Hao, C., & Wang, C. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3697846
“To address this, researchers have explored two mainstream strategies: numerical approximation [22] and model compression techniques [13].”
A QNN-Based Framework for Enhancing Reliability in Critical 5G Networks
Tiwari, A., Molakatala, N., & Rehman, A. S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-26349-0_11
fastiRNN: Fast inference of recurrent neural networks at the post-training stage
Wang, H., Chen, J., Chen, Y., & Li, J. C. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.134285
PIM-NoC: A NoC Architecture with In-Router Processing-in-Memory for DNN Acceleration
Liu, Y., Ren, H., Han, L., Qi, C., Chen, H. H., Fu, Y., & Li, L. (2026) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas66217.2026.11562908
“With the rapid advancement of artificial intelligence, Deep Neural Networks (DNNs) have achieved remarkable performance in image recognition, natural language processing, and autonomous driving, becoming a key driver of the information industry [1]–[4].”
Enabling Edge IoT Intelligence: A Survey on Resource-Constrained Continual Learning
Yang, R., Bie, D., Li, D., Yao, Z., & Xiang, J. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3705393
“Unlike cloud servers, edge IoT devices operate under the “rigid triangle constraint” of limited memory, restricted computation, and battery-dependent energy [5].”
Semi-Supervised Adversarial Learning Framework for Controller Area Network Bus Intrusion Detection
Kim, J., Im, H., Kim, S., & Lee, S. (2026) · Sensors · doi:10.3390/s26123964
Using Adaptive Activation Functions in Pre-Trained Artificial Neural Network Models
Bodyanskiy, Y. V., & Kostiuk, S. (2023) · International Conference on Information Control Systems & Technologies (ICST) · doi:10.55056/ceur-ws.org/vol-3513/paper8.pdf · Cited by 2
“[25] overviews model compression methods for efficient deployment to resource-constrained devices and specialized hardware accelerators.”
BackgroundMethodTsmots, I., Teslyuk, V., Opotyak, Y., & Shtohrinets, B. V. (2026) · Eastern-European Journal of Enterprise Technologies · doi:10.15587/1729-4061.2026.357361
Sparse Spatial-Temporal Fusion Model for Flight Control System Real-Time Anomaly Detection
Yang, T., Song, Y., Carratù, M., Gallo, V., Pietrosanto, A., Peng, Y., & Liu, D. M. (2026) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2026.3706760
Vesa, C., Bârsan, G., & Mănescu, B. A. (2026) · International conference KNOWLEDGE-BASED ORGANIZATION · doi:10.2478/kbo-2026-0079
“Wearable physiological monitoring has a long lineage in civilian sports science and military human-performance research [11, 12].”
A Plug-and-Play Data-Centric Training Framework for Efficient Hyperspectral Image Classification
Li, X., Xie, H., Yan, L., Xu, M., Shen, W., Shen, H., Song, J., Guo, W., & Wei, Y. (2026) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2026.3712611
“Conversely, to alleviate this computational bottleneck, a parallel line of research heavily focuses on model acceleration, predominantly adopting a “reduction-based” philosophy [63], [64].”
Realizable N:M Sparse Transformer Inference via Search–Kernel Co-design
Liu, Y., Lou, W., Wang, Z., Ke, Z., Zuo, F., Wang, C., & Zhou, X. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-35248-4_26
“Model compression, especially pruning, has been widely studied to mitigate this burden by removing redundant parameters while preserving accuracy [2].”
尚援朝, Liu, X., & Zan, F. (2026) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2026.084688
RELOAF: A Reinforcement Learning-based Online Adaptive Filtering Framework for Data Prefetching
Fu, L., Jiang, H., Jia, A., Zou, P., Ren, Z., Chen, Y., & Qiao, L. (2026) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3828525
“To solve this bottleneck, prefetchers have been proposed [2].”
OPENCLAW-SRT: Resource-Aware Autonomous Agent Architecture for Smart City Edge Computing
Chen, Y., & Luo, J. (2026) · International Conference on Smart City and Sustainable Development · doi:10.1145/3821835.3821917
Kim, J., Lee, H., Choi, Y., Jang, W. Y., Kim, H., Ponnusamy, K. M., Lee, H., Yoo, J., Lee, J. Y., Chandramohan, S., et al. (2026) · Advanced Functional Materials · doi:10.1002/adfm.77012
“However, developing fully functional hardware neural networks (HW-NNs) remains challenging [14].”
Deep Neural Network Compression via Data-Driven Low-Rank Singular Value Decomposition
Toulaoui, A., Khalfi, H., & Hafidi, I. (2026) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.18694
TinyShareNet: Toward Extremely Tiny Network for Memory Efficient On-Chip Inference at the Edge
Wu, T., Guan, J., Lai, R., Chai, C., & Zhu, Z. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3716411
“Being an e ff ective strategy in e ffi cient model design, parameter sharing aims to reuse the same weights of convolution filters from di ff erent layers, which has been extensively studied in recent years [28], [29].”
Low-Latency Industrial ML Models for Real-Time Digital Twin Synchronization
Malhotra, R., & Sn, B. (2020) · International Journal of Machine Learning and Predictive Analytics · doi:10.67228/3142788x/ijmlpa-2020pi1v4t
Lightweight vision transformer for adversarially robust object detection in autonomous vehicles
Okul, Ş. (2026) · Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi · doi:10.17714/gumusfenbil.1866586
Current Status, Challenges, and Prospects for Machine Vision Empowered by Edge Intelligence
Zhao, B., Wang, Y., Zhang, H., Wang, H., Tang, K., Liao, X., Chen, Y., Su, X., & Yin, A. (2026) · Engineering · doi:10.1016/j.eng.2026.07.017
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.