The DOI matches its Crossref record; 144 of the 157 references it cites were found in a registry or an index.
Efficient Processing of Deep Neural Networks: A Tutorial and Survey
Abstract
Deep neural networks (DNNs) are currently widely used for many artificial intelligence (AI) applications including computer vision, speech recognition, and robotics. While DNNs deliver state-of-the-art accuracy on many AI tasks, it comes at the cost of high computational complexity. Accordingly, techniques that enable efficient processing of DNNs to improve energy efficiency and throughput without sacrificing application accuracy or increasing hardware cost are critical to the wide deployment of DNNs in AI systems. This article aims to provide a comprehensive tutorial and survey about the recent advances toward the goal of enabling efficient processing of DNNs. Specifically, it will provide an overview of DNNs, discuss various hardware platforms and architectures that support DNNs, and highlight key trends in reducing the computation cost of DNNs either solely via hardware design changes or via joint hardware design and DNN algorithm changes. It will also summarize various development resources that enable researchers and practitioners to quickly get started in this field, and highlight important benchmarking metrics and design considerations that should be used for evaluating the rapidly growing number of DNN hardware designs, optionally including algorithmic codesigns, being proposed in academia and industry. The reader will take away the following concepts from this article: understand the key design considerations for DNNs; be able to evaluate different DNN hardware implementations with benchmarks and comparison metrics; understand the tradeoffs between various hardware architectures and platforms; be able to evaluate the utility of various DNN design techniques for efficient processing; and understand recent implementation 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
“Classical arithmetic operations, such as MAC, constitute the bulk of the computational requirements for models used in artificial intelligence [13].”
From Brain-inspired computing systems: a systematic literature review (2024) · cites this paper as [13]
In the citing paper
“To clarify, consider the multiply-and-accumulate (MAC) operation in digital processors, where the operation energy consumption is significantly dependent on the bit width of the input operands [13, 235] or equivalently to the signal-to-noise ratio (SNR) in analog computing [236].”
From Brain-inspired computing systems: a systematic literature review (2024) · cites this paper as [13]
In the citing paper
“Altogether, processing one 224 × 224 input image needs a VGG-16 model of 15.5 G multiply-and-accumulates and 138 M weights [40].”
From Face Recognition Based on Deep Learning and FPGA for Ethnicity Identification (2022) · cites this paper as [40]
In the citing paper
“The energy consumption of executing a DNN model is highly dictated by the computation architecture and methods (e.g., ultra-low power compute-in-memory AI accelerator) at the edge computation nodes [249], the architecture of DNN models [250], and the wireless transmission for data exchange during the model inference procedure.”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [250]
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 [25]
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 15×
“In 2015, the ImageNet winning entry, ResNet [15], exceeded human-level accuracy with a top-5 error rate4 below 5%.”
InfluentialVery 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 10×
“Used in VGG-16. decompose 5x5 filter 5x1 filter 1x5 filter Apply sequentially (b) Constructing a 5×5 support from 1×5 and 5×1 filter.”
InfluentialImageNet 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 10×
“this, local response normalization (LRN) [3] was used, which”
BackgroundMethodInfluentialHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“In contrast, recurrent neural networks (RNNs), of which Long Short-Term Memory networks (LSTMs) [38] are a popular variant, have internal memory to allow long-term dependencies to affect the output.”
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
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 · cited 4×
“Table 2 Summary of Popular DNNs [3], [15], [48], [50], [51]”
InfluentialGoing 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 · cited 8×
“†ACCURACY IS MEASURED BASED ON TOP-5 ERROR ON IMAGENET [14].”
InfluentialFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834
Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.81 · Cited by 32,684
Research works printed but not linked: 34
- 10.13140/RG.2.2.18893.74727 · DOI · Paper page
- Li, Stanford CS Class CS231n: Convolutional Neural Networks for Visual Recognition
- 2016, Complete Visual Networking Index (VNI) Forecast
- Woodhouse, 2016, Big, Big, Big Data: Higher and Higher Resolution Video Surveillance
- Maas, Rectifier nonlinearities improve neural network acoustic models, 1
- Caffe LeNet MNIST
- Caffe Model Zoo
- Matconvnet Pretrained Models
- TensorFlow-Slim Image Classification Library
- Deep Learning Frameworks
- Yann, The MNIST DATABASE of Handwritten Digits
- Krizhevsky, The CIFAR-10 Dataset
- Krizhevsky, 2010, Convolutional deep belief networks on cifar-10, Unpublished Manuscript, 40, 1
- Pascal VOC Data Sets
- Microsoft Common Objects in Context (COCO) Dataset
- Google Open Images
- YouTube-8M
- AudioSet
- 2017, Intel Architecture Instruction Set Extensions Programming Reference
- Condon, Facebook unveils Big Basin, new server geared for deep learning
- Intel Math Kernel Library
- 10.1145/1816038.1815993 · DOI
- 10.1145/2872887.2750389 · DOI
- 10.1145/2644865.2541967 · DOI · Paper page
- Standard, High bandwidth memory (HBM) DRAM
- 10.1145/3093336.3037702 · DOI
- Wilson, International technology roadmap for semiconductors (ITRS)
- 10.1109/icvd.2004.1260946 · doi:10.1109/icvd.2004.1260946
- Lee, A 2.5 GHz 7.7 TOPS/W switched-capacitor matrix multiplier with co-designed local memory in 40 nm, 418
- Lee, LogNet: Energy-efficient neural networks using logrithmic computations, 5900
- DNN Energy Estimation
- 10.1145/3140659.3080254 · DOI · Paper page
- Iandola, SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1 MB model size
- Benchmarking DNN Processors
Cited by
Items 401–500 of 1,011
Zhang, H., Chen, D., & Ko, S. (2019) · IEEE Transactions on Computers · doi:10.1109/tc.2019.2936192 · Cited by 65
“Only convolutional layers in these neural networks are considered because most of the neural network computations are in convolution layers [2].”
BackgroundSharma, M. K., Zappone, A., Assaad, M., Debbah, M. A., & Vassilaras, S. (2019) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2019.2949589 · Cited by 66
Cluster Pruning: An Efficient Filter Pruning Method for Edge AI Vision Applications
Gamanayake, C., Jayasinghe, L., Ng, B. K. K., & Yuen, C. (2020) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2020.2971418 · Cited by 69
“2, as mentioned in [49].”
A survey on Adversarial Attacks and Defenses in TextPreprint
Wang, W.-Q., Wang, L., Tang, B.-X., Wang, R., & Ye, A.-S. (2019) · arXiv.org · doi:10.48550/arxiv.1902.07285 · Cited by 43
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].”
BackgroundMultistage transfer learning for medical images
Ayana, G., Dese, K., Abagaro, A. M., Jeong, K. C., Yoon, S., & Choe, S. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10855-7 · Cited by 78
“This is particularly advantageous for medical imaging tasks where resource constraints are common (Sze et al. 2017).”
Comprehensive Evaluation of Supply Voltage Underscaling in FPGA on-Chip Memories
Salami, B., Ünsal, O., & Kestelman, A. C. (2018) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1109/micro.2018.00064 · Cited by 57
“techniques are customized for different underlying platforms such as CPUs, GPUs, FPGAs, and ASICs, as in detail surveyed in [57].”
BackgroundBenchmarking Contemporary Deep Learning Hardware and Frameworks: A Survey of Qualitative Metrics
Dai, W., & Berleant, D. (2019) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi48466.2019.00029 · Cited by 66
“Weight stationary dataflow, output stationary dataflow, no local reuse dataflow, and row stationary dataflow were developed for decreasing energy consumption of FPGAs and ASICs [21].”
InfluentialA Survey of AI Accelerators for Edge Environment
Li, W., & Liewig, M. (2020) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-45691-7_4 · Cited by 70
Zhao, H., Wang, C., Guo, R., Rong, X., Guo, J., Yang, Q., Yang, L., Zhao, Y., & Li, Y. (2022) · High Voltage · doi:10.1049/hve2.12221 · Cited by 59
AI hardware acceleration with analog memory: Microarchitectures for low energy at high speed
Chang, H.-Y., Narayanan, P., Lewis, S. C., Farinha, N. C. P., Hosokawa, K., Mackin, C., Tsai, H., Ambrogio, S., Chen, A., & Burr, G. W. (2019) · IBM Journal of Research and Development · doi:10.1147/jrd.2019.2934050 · Cited by 62
Maji, S., Banerjee, U., & Chandrakasan, A. P. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3061314 · Cited by 62
Recent 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].”
BackgroundSarno, L. D., Caroselli, A., Tonin, G., Graglia, B., Pansini, V., Causio, F. A., Gatto, A., & Chiaretti, A. (2024) · Biomedicines · doi:10.3390/biomedicines12061220 · Cited by 67
“An ANN with numerous layers between the input and output layers which is capable of learning high-level features and requires high computational power [55].”
Yang, D., Meng, W., Wang, Z., Yu, T., Li, C., Zhang, Q., Zhang, Z., Li, H., Lin, Y., Xue, F., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-65691-2 · Cited by 58
Variational Network Quantization
Achterhold, J., Koehler, J. M., Schmeink, A., & Genewein, T. (2018) · International Conference on Learning Representations · Cited by 60
Yin, S., Ouyang, P., Yang, J., Lu, T., Li, X., Liu, L., & Wei, S. (2018) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2018.2881913 · Cited by 60
“Index-based encoding technique, such as run-length compression encoding [25] and zero-free neuron array format [29], is adopted to reduce the memory storage and access for full-precision DNNs.”
BackgroundMethodInfluentialEdge deep learning for neural implants: a case study of seizure detection and prediction
Liu, X., & Richardson, A. G. (2021) · Journal of Neural Engineering · doi:10.1088/1741-2552/abf473 · Cited by 62
“While DL model complexity has been defined in different ways, it is specifically the speed of inference, along with memory requirements, that is critical for resource-restrained edge devices [59].”
PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning InferencePreprint
Ankit, A., Hajj, I. E., Chalamalasetti, S. R., Ndu, G., Foltín, M., Williams, R. S., Faraboschi, P., Hwu, W., Strachan, J. P., Roy, K., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1901.10351 · Cited by 73
“[103] provide a thorough survey of deep learning accelerators.”
BackgroundJebali, F., Majumdar, A., Turck, C., Harabi, K.-E., Faye, M.-C., Muhr, E., Walder, J.-P., Bilousov, O., Michaud, A., Vianello, E., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-44766-6 · Cited by 58
An Experimental Study of Reduced-Voltage Operation in Modern FPGAs for Neural Network Acceleration
Salami, B., Onural, E. B., Yüksel, İ. E., Koç, F., Ergin, O., Kestelman, A. C., Ünsal, O., Sarbazi‐Azad, H., & Mutlu, O. (2020) · International Conference on Dependable Systems & Networks/Proceedings · doi:10.1109/dsn48063.2020.00032 · Cited by 67
“Typically, hardware acceleration using Graphics Processing Units (GPUs) [135], Field Programmable Gate Arrays (FPGAs) [85, 101], or Application-Specific Integrated Circuits (ASICs) [22, 42, 102] leads to a significant reduction in CNN power consumption [109].”
BackgroundMethodDeep Recurrent Neural Networks Approach to Sedimentary Facies Classification Using Well Logs
dos Santos, D. T., Roisenberg, M., & Nascimento, M. D. S. (2021) · IEEE Geoscience and Remote Sensing Letters · doi:10.1109/lgrs.2021.3053383 · Cited by 56
Silicon-based optoelectronics for general-purpose matrix computation: a review
Xu, P., & Zhou, Z. (2022) · Advanced Photonics · doi:10.1117/1.ap.4.4.044001 · Cited by 60
Compressive strength analysis of fly ash-based geopolymer concrete using machine learning approaches
Emarah, D. A. (2022) · Results in Materials · doi:10.1016/j.rinma.2022.100347 · Cited by 68
Design Considerations for Efficient Deep Neural Networks on Processing-in-Memory Accelerators
Yang, T.-J., & Sze, V. (2019) · 2019 IEEE International Electron Devices Meeting (IEDM) · doi:10.1109/iedm19573.2019.8993662 · Cited by 55
“peripheral circuits are much larger than the memory array; the array size of digital accelerators typically range from 16× 16 to 128 × 128 [6] compared with 128 × 128 to 4096 × 4096 for PIM accelerators [2, 4].”
BackgroundSelf-citationA Parallel Multibit Programing Scheme With High Precision for RRAM-Based Neuromorphic Systems
Chen, J.-R., Wu, H., Gao, B., Tang, J., Hu, X. S., & Qian, H. (2020) · IEEE Transactions on Electron Devices · doi:10.1109/ted.2020.2979606 · Cited by 55
“For example, to program VGG-16 with 4-bit synaptic weights (138 M [3], ∼40 write-verify cycles for 4-bit precision [7]), it takes 46 min employing the existing technology [5] (500 ns per write-verify cycle).”
MethodCompacting Deep Neural Networks for Internet of Things: Methods and Applications
Zhang, K., Ying, H., Dai, H., Li, L., Peng, Y., Guo, K., & Yu, H. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3063497 · Cited by 69
“The success of DL is mainly attributed to ANNs, which consist of a number of bionic neurons similar to millions of neurons in a human brain [9].”
BackgroundiGniter: Interference-Aware GPU Resource Provisioning for Predictable DNN Inference in the Cloud
Xu, F., Xu, J., Chen, J., Chen, L., Shang, R., Zhou, Z., & Liu, F. (2022) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2022.3232715 · Cited by 62
“WITH the proliferating artificial intelligence applications, deep neural network (DNN) inference workloads are becoming increasingly commonplace in cloud datacenters [1].”
BackgroundRecent Research for HZO-Based Ferroelectric Memory towards In-Memory Computing Applications
Yoo, J., Song, H., Lee, H., Lim, S., Kim, S., Heo, K., & Bae, H. (2023) · Electronics · doi:10.3390/electronics12102297 · Cited by 59
Interpretability of deep neural networks: A review of methods, classification and hardware
Antamis, T., Drosou, A., Vafeiadis, T., Nizamis, A., Ioannidis, D., & Tzovaras, D. K. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128204 · Cited by 70
Zheng, S., Ouyang, P., Song, D., Li, X., Liu, L., Wei, S., & Yin, S. (2019) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2019.2942092 · Cited by 54
“5, which operates as an output-stationary dataflow [22].”
BackgroundImperceptible Misclassification Attack on Deep Learning Accelerator by Glitch Injection
Liu, W., Chang, C.-H., Zhang, F., & Lou, X. (2020) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac18072.2020.9218577 · Cited by 57
“Due to the high demand on computing power of DNN, efficient hardware accelerators have been developed to enable deep learning on resource and energy-constrained IoT endpoint devices [1].”
BackgroundStructured Probabilistic Pruning for Convolutional Neural Network AccelerationPreprint
Wang, H., Zhang, Q., Wang, Y., & Hu, H. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1709.06994 · Cited by 68
I Know What You See: Power Side-Channel Attack on Convolutional Neural Network AcceleratorsPreprint
Wei, L., Luo, B., Li, Y., Liu, Y., & Xu, Q. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.05847 · Cited by 54
Face Recognition Based on Deep Learning and FPGA for Ethnicity Identification
AlBdairi, A. J. A., Xiao, Z., Alkhayyat, A., Humaidi, A. J., Fadhel, M. A., Taher, B. H., Alzubaidi, L., Santamaría, J., & Al-Shamma, O. (2022) · Applied Sciences · doi:10.3390/app12052605 · Cited by 54
Memory Trojan Attack on Neural Network Accelerators
Zhao, Y. K., Hu, X., Li, S., Ye, J., Deng, L., Ji, Y., Xu, J., Wu, D., & Xie, Y. (2019) · Design, Automation, and Test in Europe · doi:10.23919/date.2019.8715027 · Cited by 69
“The development of Deep Neural Networks (DNNs) has achieved extraordinary accuracy for tasks in computer vision, speech recognition, and machine translation [1].”
BackgroundDeep learning methods for protein function prediction
Boadu, F., Lee, A., & Cheng, J. (2025) · PROTEOMICS · doi:10.1002/pmic.202300471 · Cited by 53
“Several deep learning models that have demonstrated effectiveness for dealing with sequential data are: 1) convolutional neural networks(CNNs) [22], 2) recurrent neural networks(RNNs) [23, 24], 3) deep neural networks(DNNs) [25, 26], and 4) attention-based transformers [2, 27].”
BackgroundAn overview of critical applications of resistive random access memory
Zahoor, F., Nisar, A., Bature, U. I., Abbas, H., Bashir, F. I., Chattopadhyay, A., Kaushik, B. K., Alzahrani, A., & Hussin, F. A. (2024) · Nanoscale Advances · doi:10.1039/d4na00158c · Cited by 60
Everything Leaves Footprints: Hardware Accelerated Intermittent Deep Inference
Kang, C.-K., Mendis, H. R., Lin, C., Chen⋆, M.-S., & Hsiu, P.-C. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.3012217 · Cited by 53
“Nested loops are a common processing structure in a neural network implementation [43].”
BackgroundMethodInfluentialOn Removing Algorithmic Priority Inversion from Mission-critical Machine Inference Pipelines
Liu, S., Yao, S., Fu, X., Tabish, R., Yu, S., Bansal, A., Yun, H., Sha, L., & Abdelzaher, T. (2020) · IEEE Real-Time Systems Symposium (RTSS) · doi:10.1109/rtss49844.2020.00037 · Cited by 69
Yan, Y., Stewart, T. C., Choo, X., Vogginger, B., Partzsch, J., Höppner, S., Kelber, F., Eliasmith, C., Furber, S. B., & Mayr, C. (2021) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/abf150 · Cited by 63
“The neural networks employed in these AI applications are most commonly deep neural networks (DNNs).”
MethodInfluentialLiu, Z., Ren, E., Qiao, F., Wei, Q., Liu, X.-J., Luo, L., Zhao, H., & Yang, H. (2020) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2020.2984161 · Cited by 54
“Previous works [18], [19] have pointed out the analog readout bottleneck existing in CMOS image sensors, and [5], [20] have emphasized the overhead of data conversion lying in conventional CIM architectures.”
BackgroundPhysical Side-Channel Attacks on Embedded Neural Networks: A Survey
Real, M. M., & Salvador, R. (2021) · Applied Sciences · doi:10.3390/app11156790 · Cited by 75
“Further details on CNNs can be found in [9,11].”
BackgroundEvaluation Method of Deep Learning-Based Embedded Systems for Traffic Sign Detection
Lopez-Montiel, M. A., Orozco-Rosas, U., Sánchez-Adame, M., Picos, K., & Montiel, O. (2021) · IEEE Access · doi:10.1109/access.2021.3097969 · Cited by 61
“Most DNN models are based on 32-bit floating-point matrix multiplication [22].”
BackgroundMethodInfluentialDesign Space Exploration of Hardware Spiking Neurons for Embedded Artificial Intelligence
Abderrahmane, N., Lemaire, E., & Miramond, B. (2019) · Neural Networks · doi:10.1016/j.neunet.2019.09.024 · Cited by 56
“In this paper, we are dealing with supervised feed-forward networks trained with back-propagation learning, because they are the most dominant deployedmodels when considering hardware integration[41][10].”
MethodReliable customer analysis using federated learning and exploring deep-attention edge intelligence
Ahmed, U., Srivastava, G., & Lin, J. C.-W. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2021.08.028 · Cited by 52
Dinelli, G., Meoni, G., Rapuano, E., Benelli, G., & Fanucci, L. (2019) · International Journal of Reconfigurable Computing · doi:10.1155/2019/7218758 · Cited by 60
“'e main advantage offered by a fixed-point representation is the possibility to shrink the model dimension and complexity with a negligible loss in accuracy [22].”
BackgroundWang, X., Hua, H., & Xu, Y. (2020) · IEEE Access · doi:10.1109/access.2020.2978253 · Cited by 61
“Neural network is inspired by the concept that a neuron’s computation involves a weighted sum of the input values [39], and the typical DNN model can be seen as a multi-layer perceptron (MLP).”
BackgroundZeinolabedin, S. M. A., Schüffny, F. M., George, R., Kelber, F., Bauer, H., Scholze, S., Hänzsche, S., Stolba, M., Dixius, A., Ellguth, G., et al. (2022) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2022.3142987 · Cited by 53
Deep Learning in Agriculture: A Review
Bharman, P., Saad, S. A., Khan, S., Jahan, I., Ray, M., & Biswas, M. (2022) · Asian Journal of Research in Computer Science · doi:10.9734/ajrcos/2022/v13i230311 · Cited by 60
“In next study [164], the authors suggested an optimized DNN using Jaya Algorithm to identify and classify diseases of paddy leaves.”
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
“Some researchers focus on reducing the on-board inference latency by DNN model compression [10] or DNN hardware accelerators [11].”
Recent Progress in Neuromorphic Computing from Memristive Devices to Neuromorphic Chips
Xiao, Y., Gao, C., Jin, J., Sun, W., Wang, B., Bao, Y., Liu, C., Huang, W., Zeng, H., & Yu, Y. (2024) · Advanced Devices & Instrumentation · doi:10.34133/adi.0044 · Cited by 68
Spatial Correlation and Value Prediction in Convolutional Neural Networks
Shomron, G., & Weiser, U. C. (2018) · IEEE Computer Architecture Letters · doi:10.1109/lca.2018.2890236 · Cited by 56
Deep Learning‐Based Single‐Cell Optical Image Studies
Sun, J., Tárnok, A., & Su, X. (2020) · Cytometry Part A · doi:10.1002/cyto.a.23973 · Cited by 55
“The outstanding performance of DNN stems from its ability to extract high-level features from raw data via statistical learning methods (56), which is different from previous methods of manually extracting features or expert design rules.”
MethodShiftCNN: Generalized Low-Precision Architecture for Inference of Convolutional Neural NetworksPreprint
Gudovskiy, D., & Rigazio, L. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1706.02393 · Cited by 73
“Today numerous research and implementation efforts (Sze et al., 2017) are conducted in the area of computationally inexpensive and power-efficient CNN architectures and their implementation for embedded systems as well as data center deployments.”
BackgroundLee, 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 68
“Therefore, minimizing memory access cost is the main consideration for energy savings for CNNs [155].”
BackgroundMethodInfluentialQazi, S., Kadri, M. B., Naveed, M., Khawaja, B. A., Khan, S. Z., Alam, M. M., & Su’ud, M. M. (2024) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2024.048893 · Cited by 101
Neuromorphic computing's yesterday, today, and tomorrow – an evolutional view
Chen, Y., Li, H. H., Wu, C., Song, C., Li, S. K., Min, C., Cheng, H.-P., Wen, W., & Liu, X. (2017) · Integration · doi:10.1016/j.vlsi.2017.11.001 · Cited by 51
Self-citationLee, J.-H., Lim, D., Jeong, H., Ma, H., & Shi, L. (2019) · IEEE Transactions on Electron Devices · doi:10.1109/ted.2019.2906249 · Cited by 58
“ARTIFICIAL neural networks (ANNs) have achieved unprecedented performance in various domains, such as object recognition and speech recognition [1], [2].”
BackgroundImprove Image Classification Using Data Augmentation and Neural Networks
Gu, S.-Q., Pednekar, M., & Slater, R. (2019) · SMU Data Science Review · Cited by 75
Pei, L., Xia, S., Chu, L., Xiao, F., Wu, Q., Yu, W., & Qiu, R. C. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3055859 · Cited by 52
“In this work, following the prior work [41] and considering the characteristics of IMU data samples, the large stride size is employed, instead of the pooling strategies, to persevere the spatial structure of the considered data.”
MethodNeural Architecture Search and Hardware Accelerator Co-Search: A Survey
Sekanina, L. (2021) · IEEE Access · doi:10.1109/access.2021.3126685 · Cited by 60
“In the case of ResNet-50, the ratio of MACs conducted in convolutional to fully connected layers is 1930× [9].”
BackgroundInfluentialArunan, A., Sirojan, T., Ravishankar, J., & Eliathamby Ambikairajah, A. (2020) · IEEE Systems Journal · doi:10.1109/jsyst.2020.2986577 · Cited by 49
“Implementing these DNN algorithms in a real-time embedded device is much harder due to limited memory and presence of comparatively less efficient processors in the embedded devices [35].”
Remote Power Side-Channel Attacks on BNN Accelerators in FPGAs
Moini, S., Tian, S., Holcomb, D. E., Szefer, J. M., & Tessier, R. (2021) · Design, Automation, and Test in Europe · doi:10.23919/date51398.2021.9473915 · Cited by 49
An Overview of Energy-Efficient Hardware Accelerators for On-Device Deep-Neural-Network Training
Lee, J., & Yoo, H. (2021) · IEEE Open Journal of the Solid-State Circuits Society · doi:10.1109/ojsscs.2021.3119554 · Cited by 57
FitNN: A Low-Resource FPGA-Based CNN Accelerator for Drones
Zhang, Z., Mahmud, M. A. P., & Kouzani, A. Z. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3179016 · Cited by 57
“The challenge is to implement intensive parameters and computations of the DNN on the edge device, e.g., a drone, which particularly has low computational capability due to its limited resources and battery life [11], [14].”
BackgroundInfluentialSTONNE: Enabling Cycle-Level Microarchitectural Simulation for DNN Inference Accelerators
Muñoz-Martínez, F., Abellán, J. L., Acacio, M. E., & Krishna, T. (2021) · IEEE International Symposium on Workload Characterization · doi:10.1109/iiswc53511.2021.00028 · Cited by 56
Zheng, L., Niu, J., Long, M., & Fan, Y. (2023) · British Journal of Educational Technology · doi:10.1111/bjet.13283 · Cited by 49
Gu, M., Zhang, Y., Wen, Y., Ai, G., Zhang, H., Wang, P., & Wang, G. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106623 · Cited by 55
Spiking neural networks for frame-based and event-based single object localization
Barchid, S., Mennesson, J., Eshraghian, J. K., Djéraba, C., & Bennamoun, M. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126805 · Cited by 67
Towards Energy Efficient non-von Neumann Architectures for Deep Learning
Ganguly, A., Muralidhar, R., & Singh, V. (2019) · International Symposium on Quality Electronic Design (ISQED) · doi:10.1109/isqed.2019.8697354 · Cited by 52
“for the behavioral characteristics of the neural algorithms [3].”
BackgroundBayesian convolutional neural network estimation for pediatric pneumonia detection and diagnosis
Fernandes, V., Bráz, G., de Paiva, A. C., Silva, A. C., & Gattass, M. (2021) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2021.106259 · Cited by 55
Cheng, L., Karkhanis, P., Gökbağ, B., Liu, Y., & Li, L. (2022) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1008885 · Cited by 52
Gaba, S., Budhiraja, I., Kumar, V., Garg, S., Kaddoum, G., & Hassan, M. M. (2022) · Computer Communications · doi:10.1016/j.comcom.2022.05.035 · Cited by 52
SRAM Cell Design Challenges in Modern Deep Sub-Micron Technologies: An Overview
Gul, W., Shams, M., & Al-Khalili, D. (2022) · Micromachines · doi:10.3390/mi13081332 · Cited by 48
“This study forms a foundation for the understanding of SRAM challenges before exploring SRAM-based CIM for ML.”
BackgroundGlobal Aligned Structured Sparsity Learning for Efficient Image Super-Resolution
Wang, H., Zhang, Y., Qin, C., Gool, L. V., & Fu, Y. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3268675 · Cited by 51
“ters, neural network pruning [12], [16], [26] is a promising direction.”
BackgroundInfluentialDFTerNet: Towards 2-bit Dynamic Fusion Networks for Accurate Human Activity Recognition
Yang, Z., Raymond, O. I., Zhang, C., Wan, Y., & Long, J. (2018) · IEEE Access · doi:10.1109/access.2018.2873315 · Cited by 51
“(16/32)-bit floating point, 2-bit (even 8-bit) operations will not only reduce the energy costs for Hardware/IC Design (see Figure 1), but also halves the memory access costs and memory size requirements during runtime, which will greatly facilitate the deployment of binary/ternary convolutional neural networks on por…”
BackgroundMASR: A Modular Accelerator for Sparse RNNs
Gupta, U., Reagen, B., Pentecost, L., Donato, M., Tambe, T., Rush, A. M., Wei, G.-Y., & Brooks, D. M. (2019) · International Conference on Parallel Architectures and Compilation Techniques (PACT) · doi:10.1109/pact.2019.00009 · Cited by 60
“Typically, previous work either exploits sparsity in weights or activations, but not both [10], [11], [13], [14], [22], [46], [47] leaving key performance, area, and energy savings on the table.”
BackgroundA Deep Neural Network-Based Multi-Frequency Path Loss Prediction Model from 0.8 GHz to 70 GHz
Nguyen, C. M. H., & Cheema, A. A. (2021) · Sensors · doi:10.3390/s21155100 · Cited by 58
“This is because DNNs can learn high-level features with more complexity and abstraction compared to shallow neural networks [30].”
BackgroundAlbireo: Energy-Efficient Acceleration of Convolutional Neural Networks via Silicon Photonics
Shiflett, K., Karanth, A., Bunescu, R., & Louri, A. (2021) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1109/isca52012.2021.00072 · Cited by 66
“where broadcasts incur high energy costs [30], which is often circumvented through spatial or local reuse of data [54].”
BackgroundApplications of Artificial Intelligence Methodologies to Behavioral and Social Sciences
Robila, M., & Robila, S. A. (2019) · Journal of Child and Family Studies · doi:10.1007/s10826-019-01689-x · Cited by 65
“Neural network (NN) refers to an algorithmic framework that aims to simulate a simpli fi ed version of neural connections in the brain (Sze et al. 2017).”
Techniques and applications for soccer video analysis: A survey
Cuevas, C., Quilón, D., & Garcia, N. N. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-09409-0 · Cited by 61
“CNNs have proven to be among the most competitive methods for object detection in images [98, 146].”
MethodSumbul, H. E., Wu, T. F., Li, Y., Sarwar, S. S., Koven, W., Murphy-Trotzky, E., Cai, X., Ansari, E., Morris, D. H., Liu, H., et al. (2022) · Custom Integrated Circuits Conference · doi:10.1109/cicc53496.2022.9772810 · Cited by 49
A Feed-Forward and Pattern Recognition ANN Model for Network Intrusion Detection
Iqbal, A., & Aftab, S. (2019) · International Journal of Computer Network and Information Security · doi:10.5815/ijcnis.2019.04.03 · Cited by 47
“Artificial Neural Network can learn rapidly from experiences as well as from complex nonlinear problems [29].”
BackgroundNeural Compatibility Modeling With Probabilistic Knowledge Distillation
Han, X., Song, X., Yao, Y., Xu, X.-S., & Nie, L. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2936742 · Cited by 47
“Later, the knowledge distillation framework is adopted to accelerate the training process of the neural network [18] and improve the model portability [19], [20].”
MethodBioCNN: A Hardware Inference Engine for EEG-Based Emotion Detection
Gonzalez, H. A., Muzaffar, S., Yoo, J., & Elfadel, I. M. (2020) · IEEE Access · doi:10.1109/access.2020.3012900 · Cited by 52
Toğaçar, M., Cömert, Z., & Ergen, B. (2021) · Neural Computing and Applications · doi:10.1007/s00521-021-05758-5 · Cited by 49
Wu, B., Zhao, S., Sun, G., Zhang, X., Su, Z., Zeng, C., & Liu, Z. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00220 · Cited by 44
“The superior performance of deep CNNs usually relies on large amounts of labeled training data [40].”
BackgroundFalaschetti, L., Manoni, L., Rivera, R. C. F., Pau, D., Romanazzi, G., Silvestroni, O., Tomaselli, V., & Turchetti, C. (2021) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2021.3098454 · Cited by 45
Mao, W., Li, K., Cheng, Q., Dai, L., Li, B., Xie, X., Li, H., Lin, L., & Yu, H. (2021) · IEEE Transactions on Very Large Scale Integration (VLSI) Systems · doi:10.1109/tvlsi.2021.3128435 · Cited by 52
Najmaei, S., Glasmann, A., Schroeder, M. A., Sarney, W. L., Chin, M. L., & Potrepka, D. M. (2022) · Materials Today · doi:10.1016/j.mattod.2022.08.017 · Cited by 51
Park, S., Lee, D., Kang, J., Choi, H., & Park, J. (2023) · Nature Communications · doi:10.1038/s41467-023-41991-3 · Cited by 48
“As a new computing paradigm, in-memory computing enables highly ef fi cient data-intensive computation because energy is not consumed for data migration and access 4 – 6 .”
Zhang, S., Wang, L., Zhu, S., Deng, X. G., Fu, S., Luo, C., Dong, Y., & Yan, D. (2024) · Materials & Design · doi:10.1016/j.matdes.2024.113267 · Cited by 46
From 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 63
A survey of unmanned aerial vehicles and deep learning in precision agriculture
Wang, D., Zhao, M., Li, Z., Xu, S., Wu, X., Ma, X., & Liu, X. (2024) · European Journal of Agronomy · doi:10.1016/j.eja.2024.127477 · Cited by 63
PIR-DSP: An FPGA DSP Block Architecture for Multi-precision Deep Neural Networks
Rasoulinezhad, S., Zhou, H., Wang, L., & Leong, P. H. W. (2019) · IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM) · doi:10.1109/fccm.2019.00015 · Cited by 60
“A more thorough review is available in [15].”
BackgroundAn Analog Neural Network Computing Engine Using CMOS-Compatible Charge-Trap-Transistor (CTT)
Du, Y., Du, L., Gu, X., Du, J., Wang, X. S., Hu, B., Jiang, M., Chen, X., Iyer, S. S., & Chang, M.-C. F. (2018) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2018.2859237 · Cited by 59
“Furthermore, the extensive processor-memory data movements have become the dominated energy consumption [39].”
SubFlow: A Dynamic Induced-Subgraph Strategy Toward Real-Time DNN Inference and Training
Lee, S., & Nirjon, S. (2020) · IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS) · doi:10.1109/rtas48715.2020.00-20 · Cited by 63
“Unfortunately, the current feed-forward algorithms, such as the lowering method [88] do not support the sparse-neuron-aware feed-forwarding.”
MethodEmNet: a deep integrated convolutional neural network for facial emotion recognition in the wild
Saurav, S., Saini, R., & Singh, S. K. (2021) · Applied Intelligence · doi:10.1007/s10489-020-02125-0 · Cited by 50
Xu, Z., Wang, Z., Li, J., Jin, T., Meng, X., & Gao, S. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.107536 · Cited by 54
Metadata
- Authors
- Sze, V., Chen, Y.-H., Yang, T.-J., & Emer, J. S.
- Affiliations
- Massachusetts Institute of Technology (US)
- Nvidia (United States) (US)
- Venue
- Proceedings of the IEEE
- Published
- 2017-12
- Volume
- 105
- Issue
- 12
- Pages
- 2295-2329
- ISSN
- 0018-9219 (ISSN-L)
- 1558-2256
- OpenAlex
- W2604319603
- Cited by
- 4,215
- References
- 157
- Citations by year
- 2026: 504 · 2025: 523 · 2024: 474 · 2023: 466 · 2022: 508 · 2021: 568 · 2020: 551 · 2019: 392 · 2018: 199 · 2017: 28
- FWCI
- 132.75
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Proceedings of the IEEE (publishedVersion)
- Funders
- DARPA YFA
- MIT CICS
- gift from Nvidia
- gift from Intel
- 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
- deep neural network, energy efficiency, throughput, hardware accelerators, computation cost reduction, benchmarking metrics
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.