MoDNN: Local distributed mobile computing system for Deep Neural Network
The DOI matches its Crossref record; 19 of the 20 references it cites were found in a registry or an index.
Abstract
Although Deep Neural Networks (DNN) are ubiquitously utilized in many applications, it is generally difficult to deploy DNNs on resource-constrained devices, e.g., mobile platforms. Some existing attempts mainly focus on client-server computing paradigm or DNN model compression, which require either infrastructure supports or special training phases, respectively. In this work, we propose MoDNN - a local distributed mobile computing system for DNN applications. MoDNN can partition already trained DNN models onto several mobile devices to accelerate DNN computations by alleviating device-level computing cost and memory usage. TWo model partition schemes are also designed to minimize non-parallel data delivery time, including both wakeup time and transmission time. Experimental results show that when the number of worker nodes increases from 2 to 4, MoDNN can accelerate the DNN computation by 2.17-4.28 X. Besides the parallel execution, the performance speedup also partially comes from the reduction of the data delivery time, e.g., 30.02% w.r.t. conventional 2D-grids partition.
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
“MoDNN, proposed in [177], executes DL in a local distributed mobile computing system over a Wireless Local Area Network (WLAN).Each layer of DNNs is partitioned into slices to increase parallelism and to reduce memory footprint, and these slices are executed layerby-layer.By the execution parallelism among multiple end devices, the DL computation can be significantly accelerated.”
From Convergence of Edge Computing and Deep Learning: A Comprehensive Survey (2020) · cites this paper as [177]
In the citing paper
“The experiment shows that with two to four worker nodes, MoDNN accelerates the DNN model inference by 2.17-4.28×.”
From Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing (2019) · cites this paper as [85]
Citations
Cites
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 4×
“A representative example is V GG [1], which demonstrates state-of-the-art performance in ImageNet Large Scale Visual Recognition Challenge 2014 (ILSVRC14).”
BackgroundMethodInfluentialGradient-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 1×
“1 depicts our measurement results of the computing time and the memory usage of different network layers from three popular DNN models running on smartphones – Lenet [9], Inception - BN [10], and V GG [1].”
BackgroundBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818 · cited 1×
“1 depicts our measurement results of the computing time and the memory usage of different network layers from three popular DNN models running on smartphones – Lenet [9], Inception - BN [10], and V GG [1].”
BackgroundMapReduce: simplified data processing on large clusters
Dean, J., & Ghemawat, S. (2008) · Communications of the ACM · doi:10.1145/1327452.1327492 · Cited by 18,476 · cited 1×
“MapReduce is a programming model for simplifying parallel data processing in distributed systems [7].”
BackgroundBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 14,214
A tutorial on spectral clustering
von Luxburg, U. (2007) · Statistics and Computing · doi:10.1007/s11222-007-9033-z · Cited by 10,349
Some simplified NP-complete graph problems
Garey, M. R., Johnson, D. S., & Stockmeyer, L. (1976) · Theoretical Computer Science · doi:10.1016/0304-3975(76)90059-1 · Cited by 2,059 · cited 1×
“If we consider the weight matrix as an undirected graph, generating k clusters with minimal connections between them is a NP hard problem [14].”
BackgroundMXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems
Chen, T., Li, M., Li, Y., Lin, M., Wang, N., Wang, M., Xiao, T., Xu, B., Zhang, C., & Zhang, Z. (2015) · arXiv · doi:10.48550/arxiv.1512.01274 · Cited by 1,819 · cited 2×
“We modified and recompiled the MXNet libraries so that it can support Android systems with ARM architecture through JAVA Native Interface (JNI) [17].”
MethodCo-clustering documents and words using bipartite spectral graph partitioning
Dhillon, I. S. (2001) · Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/502512.502550 · Cited by 1,697 · cited 1×
“Therefore, spectral co-clustering algorithm is introduced to address this drawback by normalizing the original connection matrix A to A norm and performing Singular Value Decomposition (SVD) on A norm [16].”
MethodScaling distributed machine learning with the parameter server
Li, M., Andersen, D. G., Park, J. W., Smola, A. J., Ahmed, A. A., Josifovski, V., Long, J. E., Shekita, E. J., & Su, B.-Y. (2014) · Operating Systems Design and Implementation · doi:10.5555/2685048.2685095 · Cited by 1,021
Research works printed but not linked: 2
- han, 2015, Deep compression: Compressing deep neural network with pruning, trained quantization and huffman coding, Computing ResearchRepository, 2
- 0, Wifi Alliance Wifi Direct Specifications
Cited by
Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing
Zhou, Z., Chen, X., Li, E., Zeng, L., Luo, K., & Zhang, J. (2019) · Proceedings of the IEEE · doi:10.1109/jproc.2019.2918951 · Cited by 2,455
“Note that DNN layers are partitioned horizontally in MoDNN and MeDNN, whereas DeepThings [86] employs a fused tile partitioning method that partitions the DNN layers vertically to reduce the memory footprint.”
MethodInfluentialSelf-citationDeep Learning in Mobile and Wireless Networking: A Survey
Zhang, C., Patras, P., & Haddadi, H. (2019) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2019.2904897 · Cited by 1,768
Deep Learning With Edge Computing: A Review
Chen, J., & Ran, X. (2019) · Proceedings of the IEEE · doi:10.1109/jproc.2019.2921977 · Cited by 1,540
“[107] J. Mao, X. Chen, K. W. Nixon, C. Krieger, and Y. Chen, “MoDNN: Local distributed mobile computing system for deep neural network,” in Proc.”
MethodInfluentialConvergence of Edge Computing and Deep Learning: A Comprehensive Survey
Wang, X., Han, Y., Leung, V. C. M., Niyato, D., Yan, X., & Chen, X. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2020.2970550 · Cited by 1,498
“MoDNN, proposed in [177], executes DL in a local distributed mobile computing system”
MethodSelf-citationEdge Machine Learning for AI-Enabled IoT Devices: A Review
Merenda, M., Porcaro, C., & Iero, D. (2020) · Sensors · doi:10.3390/s20092533 · Cited by 523
DeepThings: Distributed Adaptive Deep Learning Inference on Resource-Constrained IoT Edge Clusters
Zhao, Z., Barijough, K. M., & Gerstlauer, A. (2018) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2018.2858384 · Cited by 502
“For a BODP- and MR-based execution model, an optimal approach is thereby to match the number of partitions to the number of available devices [16], which we will use for all further comparisons.”
BackgroundMethodInfluentialCommunication-efficient edge ai: algorithms and systems
Shi, Y., Yang, K., Jiang, T., Zhang, J., & Letaief, K. B. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2020.3007787 · Cited by 488
“Different methods have been developed [207], [208] to partition a pre-trained DNN over several mobile devices in order to accelerate DNN inference on devices.”
MethodEfficient 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
“MoDNN [371] is a horizontally distributed platform for DL inference in resource-constrained edge devices.”
BackgroundA Survey of Recent Advances in Edge-Computing-Powered Artificial Intelligence of Things
Chang, Z., Liu, S., Xiong, X., Cai, Z., & Tu, G. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3088875 · Cited by 418
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
“The model partitioning includes three types: partition between servers and devices [57], [176], [177], the partition between devices horizontally [178], [179] and vertically [180].”
BackgroundSPINN: synergistic progressive inference of neural networks over device and cloud
Laskaridis, S., Venieris, S. I., Almeida, M., Leontiadis, I., & Lane, N. D. (2020) · Proceedings of the 26th Annual International Conference on Mobile Computing and Networking · doi:10.1145/3372224.3419194 · Cited by 310
“Simultaneously, various works [50, 51, 85] have examined the case where the client device can offload to other devices in the local network.”
BackgroundZeng, L., Chen, X., Zhou, Z., Yang, L., & Zhang, J. (2021) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2020.3042320 · Cited by 313
“On the one hand, the local approach does not incur communication costs, while MoDNN and Musical Chair need frequent cross-device communication during the runtime, which takes energy.”
BackgroundMethodInfluentialSelf-citationAutonomous Vehicles Enabled by the Integration of IoT, Edge Intelligence, 5G, and Blockchain
Biswas, A., & Wang, H.-C. (2023) · Sensors · doi:10.3390/s23041963 · Cited by 274
Edge Intelligence in Intelligent Transportation Systems: A Survey
Gong, T., Zhu, L., Yu, F. R., & Tang, T. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3275741 · Cited by 256
“More works of DNN partition proved that it is a promising way in the future of DNN inference, which will significantly accelerate the inference process [125], [126],”
End-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
“…involves the segmentation of neurons such as weight matrix, filters, and convolution kernels into distinct parts along specific dimensions and calculating them separately on different end-edge-cloud nodes, such as MoDNN [109] for end-end collaboration and DeepThings [110] for edge-edge collaboration.”
Distributed Inference Acceleration with Adaptive DNN Partitioning and Offloading
Mohammed, T., Joe-Wong, C., Babbar, R., & Francesco, M. D. (2020) · IEEE INFOCOM 2020 - IEEE Conference on Computer Communications · doi:10.1109/infocom41043.2020.9155237 · Cited by 211
“Accordingly, the approach in this work relies on partitioning convolutional layers into segments along the largest dimension of the input [15].”
BackgroundMethodBaccour, E., Mhaisen, N., Abdellatif, A. A., Erbad, A., Mohamed, A., Hamdi, M., & Guizani, M. (2022) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3200740 · Cited by 186
“These splitting strategies are introduced and adopted by multiple recent works, including [196], [198], [199], for which we will thoroughly review the resource management techniques in the following section.”
BackgroundMethodMachine learning techniques in emerging cloud computing integrated paradigms: A survey and taxonomy
Soni, D., & Kumar, N. (2022) · Journal of Network and Computer Applications · doi:10.1016/j.jnca.2022.103419 · Cited by 172
Towards Building The Federatedgpt: Federated Instruction Tuning
Zhang, J., Vahidian, S., Kuo, M., Li, C., Zhang, R., Yu, T., Wang, G., & Chen, Y. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447454 · Cited by 363
“Given that clients often have limited computational resources in comparison to traditional centralized training cloud servers, which can uti-6915 lize thousands of GPUs [15] to fully fine-tune all parameters of LLMs, we resort to parameter-efficient tuning techniques.”
BackgroundSelf-citationToward Collaborative Inferencing of Deep Neural Networks on Internet-of-Things Devices
Hadidi, R., Cao, J., Ryoo, M. S., & Kim, H. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.2972000 · Cited by 154
“Recently, with the prevalence of large DNN models, distributing a single model has gained the attention of researchers [8], [21]–[25].”
BackgroundCharacterizing the deployment of deep neural networks on commercial edge devices
Hadidi, R., Cao, J., Xie, Y., Asgari, B., Krishna, T., & Kim, H. (2019) · 2019 IEEE International Symposium on Workload Characterization (IISWC) · doi:10.1109/iiswc47752.2019.9041955 · Cited by 118
“MoDNN [89] creates a local distributed mobile computing system and accelerates DNN computations with model parallelism.”
Gong, L., Wang, C., Li, X., Chen, H., & Zhou, X. (2018) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2018.2857078 · Cited by 118
“This hardware underutilization is even unacceptable on mobile and embedded devices where local processing of CNN tasks is emerging [21].”
BackgroundAdaptive parallel execution of deep neural networks on heterogeneous edge devices
Zhou, L., Samavatian, M. H., Bacha, A., Majumdar, S., & Teodorescu, R. (2019) · ACM/IEEE Symposium on Edge Computing (SEC) · doi:10.1145/3318216.3363312 · Cited by 134
HyPar: Towards Hybrid Parallelism for Deep Learning Accelerator Array
Song, L., Mao, J., Zhuo, Y., Qian, X., Li, H., & Chen, Y. (2019) · 2019 IEEE International Symposium on High Performance Computer Architecture (HPCA) · doi:10.1109/hpca.2019.00027 · Cited by 108
Self-citationDeepSlicing: Collaborative and Adaptive CNN Inference With Low Latency
Zhang, S., Zhang, S., Qian, Z., Wu, J., Jin, Y., & Lu, S. (2021) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2021.3058532 · Cited by 127
“works are implemented and used as baselines: DeepThings [27] and MoDNN [26].”
BackgroundMethodInfluentialResource Management in Mobile Edge Computing: A Comprehensive Survey
Zhang, X., & Debroy, S. (2023) · ACM Computing Surveys · doi:10.1145/3589639 · Cited by 136
“MoDNN [117] is a distributed mobile computing system in Wireless Local Area Network (WLAN) where the DNN models are partitioned by layers and mapped onto mobile devices to accelerate the computation.”
BackgroundA Survey on Collaborative DNN Inference for Edge Intelligence
Ren, W., Qu, Y., Dong, C., Jing, Y., Sun, H., Wu, Q., & Guo, S. (2023) · Machine Intelligence Research · doi:10.1007/s11633-022-1391-7 · Cited by 137
“DeepSlicing[95] Computer vision Model partition Data partition Lower latency Less memory footprint”
BackgroundInfluentialSCOPE: A Stochastic Computing Engine for DRAM-Based In-Situ Accelerator
Li, S., Glova, A. O., Hu, X., Gu, P., Niu, D., Malladi, K. T., Zheng, H., Brennan, B. A., & Xie, Y. (2018) · IEEE/ACM International Symposium on Microarchitecture (MICRO) · doi:10.1109/micro.2018.00062 · Cited by 131
Edge AI for Internet of Energy: Challenges and perspectives
Himeur, Y., Sayed, A. N., Alsalemi, A., Bensaali, F., & Amira, A. (2023) · Internet of Things · doi:10.1016/j.iot.2023.101035 · Cited by 130
Zhao, T., Xie, Y., Wang, Y., Cheng, J., Guo, X., Hu, B., & Chen, Y. (2022) · Proceedings of the IEEE · doi:10.1109/jproc.2022.3153408 · Cited by 92
Self-citationMeDNN: A distributed mobile system with enhanced partition and deployment for large-scale DNNs
Mao, J., Yang, Z., Wen, W., Wu, C., Song, L., Nixon, K. W., Chen, X., Li, H., & Chen, Y. (2017) · 2017 IEEE/ACM International Conference on Computer-Aided Design (ICCAD) · doi:10.1109/iccad.2017.8203852 · Cited by 97
“Its effectiveness has been proven in many machine learning applications on mobile platforms [8][9].”
BackgroundSelf-citationDistributed Inference with Deep Learning Models across Heterogeneous Edge Devices
Hu, C., & Li, B. (2022) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom48880.2022.9796896 · Cited by 112
“MODNN [19] assigns workload according to the computation capacity.”
BackgroundAI on the edge: a comprehensive review
Su, W., Li, L., Liu, F., He, M., & Liang, X. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-022-10141-4 · Cited by 105
Distributed Perception by Collaborative Robots
Hadidi, R., Cao, J., Woodward, M., Ryoo, M. S., & Kim, H. (2018) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2018.2856261 · Cited by 97
“However, MoDNN does not consider real-time performance because its most optimized system with four Nexus-5 devices has a latency of six seconds.”
BackgroundInfluentialAccelerating DNN Inference With Reliability Guarantee in Vehicular Edge Computing
Liu, K., Liu, C., Yan, G., Lee, V. C. S., & Cao, J. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3279512 · Cited by 87
“On the other hand, many studies have been focused on the DNN model partitioning and offloading in edge computing environments to speed up the inference [11], [12], [13], [14], [15], [16].”
BackgroundMethodDeep Learning Challenges and Prospects in Wireless Sensor Network Deployment
Qiu, Y., Ma, L., & Priyadarshi, R. (2024) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-024-10079-6 · Cited by 97
“The process of unsupervised learning makes it possible for autoencoders to recognize the regular patterns that are inherent in the data [101–103].”
Edge Computing Technology Enablers: A Systematic Lecture Study
Douch, S., Abid, M. R., Zine-Dine, K., Bouzidi, D., & Benhaddou, D. (2022) · IEEE Access · doi:10.1109/access.2022.3183634 · Cited by 97
“In the case of horizontal inference distributing, one of the first proposed distributed inference frameworks is MoDNN [282], a map-reduce mechanism that makes partitions of the input feature according to the number of ENs, where each node does a part of the inference calculation, and an aggregation…”
BackgroundMoving Deep Learning to the Edge
Véstias, M. P., Duarte, R. P., de Sousa, J. T., & Neto, H. C. (2020) · Algorithms · doi:10.3390/a13050125 · Cited by 73
“Collaborative computation among end devices [175] executes a DL model in a distributed mobile computing network of devices.”
BackgroundEnabling Resource-Efficient AIoT System With Cross-Level Optimization: A Survey
Liu, S., Guo, B., Fang, C., Wang, Z., Luo, S., Zhou, Z., & Yu, Z. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2023.3319952 · Cited by 82
“For example, Modnn [86] significantly speeds up DL inference by introducing execution parallelism among multiple devices [87].”
BackgroundMethodInfluentialToward Reliable DNN-Based Task Partitioning and Offloading in Vehicular Edge Computing
Liu, C., & Liu, K. (2023) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2023.3280484 · Cited by 75
“Mao et al. [20] proposed the Biased One-Dimensional Partition (BODP) scheme to partition the input data of each layer in DNN models, which achieves layer-wise partitioning for accelerating DNN inference.”
BackgroundMethodInfluentialImproving 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
“For instance, it takes 16 seconds on mobile to complete an image recognition using VGG16 [25], which is intolerable for most applications [18].”
BackgroundAccPar: Tensor Partitioning for Heterogeneous Deep Learning Accelerators
Song, L., Chen, F., Zhuo, Y., Qian, X., Li, H. H., & Chen, Y. (2020) · IEEE International Symposium on High-Performance Computer Architecture (HPCA) · doi:10.1109/hpca47549.2020.00036 · Cited by 91
“…DNN [69], Bit Fusion [70], Cn-vlutin [99], TIE [137], Laconic [138], ADM-NN [139], Gist [140] Co-Design DPS-CNN [71], Minerva [72], MoDNN [73], MeDNN [74], AdaLearner [75], Stitch-X [76], PIM-DNN [77], Scalpel [78], CirCNN [79], CoSMIC [80], SnaPEA [81], OLAcce [82], Prediction-based DNN…”
Self-citationDistrEdge: Speeding up Convolutional Neural Network Inference on Distributed Edge Devices
Hou, X., Guan, Y., Han, T., & Zhang, N. (2022) · IEEE International Parallel and Distributed Processing Symposium (IPDPS) · doi:10.1109/ipdps53621.2022.00110 · Cited by 80
DeeperThings: Fully Distributed CNN Inference on Resource-Constrained Edge Devices
Stahl, R., Hoffman, A., Mueller-Gritschneder, D., Gerstlauer, A., & Schlichtmann, U. (2021) · International Journal of Parallel Programming · doi:10.1007/s10766-021-00712-3 · Cited by 74
“MoDNN distributes a CNN across multiple mobile phones connected via a wireless network.”
MethodJouhari, M., Al-Ali, A. K., Baccour, E., Mohamed, A. M. S., Erbad, A., Guizani, M., & Hamdi, M. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3079164 · Cited by 66
“Authors in [33] proposed a per-input partition scheme for pre-trained DNN models to fit with resource-constrained mobile devices and accelerate DNN computations.”
Low Latency Deep Learning Inference Model for Distributed Intelligent IoT Edge Clusters
Naveen, S., Kounte, M. R., & Ahmed, M. R. (2021) · IEEE Access · doi:10.1109/access.2021.3131396 · Cited by 77
“The inference latency is compared with DeepThings [36] and MODNN [37] on a single data frame.”
BackgroundMethodGNN at the Edge: Cost-Efficient Graph Neural Network Processing Over Distributed Edge Servers
Zeng, L., Yang, C., Huang, P., Zhou, Z., Yu, S., & Chen, X. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3229422 · Cited by 73
“Existing efforts on edge-enabled distributed intelligence have relied on either offloading computation with external infrastructures [20], [21], [22] or partitioning model execution across edge servers [23], [24], [25].”
BackgroundSelf-citationAdaptive Computation Partitioning and Offloading in Real-Time Sustainable Vehicular Edge Computing
Ku, Y.-J., Baidya, S., & Dey, S. (2021) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2021.3119585 · Cited by 60
“To consider task dependency during partitioning and offload-ing, [17], [18] divide the sequential convolution layers of a Deep Neural Network (DNN) into several independent subtasks.”
BackgroundMethodLegoDNN: block-grained scaling of deep neural networks for mobile vision
Han, R., Zhang, Q., Liu, C. H., Wang, G., Tang, J., & Chen, L. Y. (2021) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3447993.3483249 · Cited by 71
“output results [29, 53, 54], edge caching of inference results [6– 8], and model segmentation [22, 25, 31, 38, 39, 67].”
Memory- and Communication-Aware Model Compression for Distributed Deep Learning Inference on IoT
Bhardwaj, K., Lin, C.-Y., Sartor, A., & Marculescu, R. (2019) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3358205 · Cited by 53
“mpression Aware Inference our work Quantization [8, 12] X × × X Pruning [4, 9, 13, 23] X × × X Separable convolutions [3, 19, 26] X × × X KD [1, 6, 25] X × × × SplitNet [10] × X X × MoDNN, DeepThings [14, 28] × X X X ProposedNoNN X X X X in per node energy w.r.t. teacher. Moreover, for distributed inference on multipl…”
BackgroundMethodInfluentialAutomated Exploration and Implementation of Distributed CNN Inference at the Edge
Guo, X., Pimentel, A. D., & Stefanov, T. P. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3237572 · Cited by 51
“For example, MoDNN [22] splits convolution layers and fully connected layers in the VGG-16 model.”
PipeEdge: Pipeline Parallelism for Large-Scale Model Inference on Heterogeneous Edge Devices
Hu, Y., Imes, C., Zhao, X., Kundu, S., Beerel, P. A., Crago, S. P., & Walters, J. P. (2022) · Euromicro Conference on Digital System Design (DSD) · doi:10.1109/dsd57027.2022.00048 · Cited by 75
“Considering the limited bandwidth and uncertain delay between the edge and the cloud, MoDNN [63]”
BackgroundFully Distributed Deep Learning Inference on Resource-Constrained Edge Devices
Stahl, R., Zhao, Z., Mueller-Gritschneder, D., Gerstlauer, A., & Schlichtmann, U. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-27562-4_6 · Cited by 60
“Other approaches [6] are able to distribute an entire network, but have limitations in the type and size of layers that can be fitted, and they do not consider opportunities to improve communication demands further.”
BackgroundMethodInfluentialA Lightweight Collaborative Deep Neural Network for the Mobile Web in Edge Cloud
Huang, Y., Qiao, X., Ren, P., Liu, L., Pu, C., Dustdar, S., & Chen, J. (2020) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2020.3043051 · Cited by 53
“MoDNN [45] proposes two partition schemes to minimize non-parallel data delivery latency and accelerate DNN inference by alleviating device-level computing cost and memory usage.”
MethodTPrune: Efficient Transformer Pruning for Mobile Devices
Mao, J., Yang, H., Li, A., Li, H., & Chen, Y. (2021) · ACM Transactions on Cyber-Physical Systems · doi:10.1145/3446640 · Cited by 42
“To understand the properties of different types of layers in deep neural networks, we deploy three representative models on Nexus5 [6, 31], which are Lenet [32], Inception-BN [29], and VGG16 [33].”
BackgroundMethodInfluentialSelf-citationDai, P., Han, B., Li, K., Xu, X., Xing, H., & Liu, K. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3457793 · Cited by 54
“As a response to this issue, multi-device cooperative DNN inference frameworks [6], [7] have been identified as a promising solution.”
Model Parallelism Optimization for Distributed Inference via Decoupled CNN Structure
Du, J., Zhu, X., Shen, M., Du, Y., Lu, Y., Xiao, N., & Liao, X. (2020) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2020.3041474 · Cited by 44
“To eliminate the impact of Scheme-1, we provide a fair comparison of DeCNN(Scheme-1,2) with other two distributed approaches such as MoDNN [6] and Coates et al. [16] using decoupled models.”
BackgroundMethodResultInfluentialBaccarelli, E., Scarpiniti, M., Momenzadeh, A., & Ahrabi, S. S. (2021) · IEEE Access · doi:10.1109/access.2021.3058021 · Cited by 42
“Specifically, user mobility is the focus of the MoDNN in [23], which proposes a distributed mobile computing system for the parallel execution of DNN applications.”
BackgroundMethodDynO: Dynamic Onloading of Deep Neural Networks from Cloud to Device
Almeida, M., Laskaridis, S., Venieris, S. I., Leontiadis, I., & Lane, N. D. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3510831 · Cited by 52
“A diferent set of related work concerns oloading CNN inference to devices in the local network [42, 43, 65], to multiple servers for load balancing of denser tasks [64], to third-party edge devices along with the CNN computation graph [29], or to either device or cloud through model selection [17].”
BackgroundJoint multi-user DNN partitioning and task offloading in mobile edge computing
Liao, Z., Hu, W., Huang, J., & Wang, J. (2023) · Ad Hoc Networks · doi:10.1016/j.adhoc.2023.103156 · Cited by 42
CrowdHMT: Crowd Intelligence With the Deep Fusion of Human, Machine, and IoT
Guo, B., Liu, Y., Liu, S., Yu, Z., & Zhou, X. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3194726 · Cited by 38
Distributed Assignment With Load Balancing for DNN Inference at the Edge
Xu, Y., Mohammed, T., Francesco, M. D., & Fischione, C. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3205410 · Cited by 43
“MoDNN [49] considers mobile devices connected through high-rate WiFi links forming a local cluster.”
BackgroundEFFECT-DNN: Energy-efficient Edge Framework for Real-time DNN Inference
Zhang, X., Mounesan, M., & Debroy, S. (2023) · IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM) · doi:10.1109/wowmom57956.2023.00015 · Cited by 42
“MoDNN [17] partitions the DNN in layers and then maps different parts of a layer onto mobile devices to accelerate the computations.”
MethodScission: Performance-driven and Context-aware Cloud-Edge Distribution of Deep Neural Networks
Lockhart, L., Harvey, P., Imai, P., Willis, P., & Varghese, B. (2020) · IEEE/ACM International Conference on Utility and Cloud Computing (UCC) · doi:10.1109/ucc48980.2020.00044 · Cited by 43
“Examples include DeepThings [33] and MoDNN [34].”
MethodYe, S., Zeng, L., Chu, X., Xing, G., & Chen, X. (2024) · ACM Annual International Conference on Mobile Computing and Networking (MobiCom) · doi:10.1145/3636534.3649363 · Cited by 64
“CoEdge, DeepThings and MoDNN [36, 62, 64] distributed execution of CNN-based inference applications on resource-constrained edge clusters.”
BackgroundSelf-citationAn Efficient Mapping Approach to Large-Scale DNNs on Multi-FPGA Architectures
Zhang, W., Zhang, J., Shen, M., Luo, G., & Xiao, N. (2019) · Design, Automation, and Test in Europe · doi:10.23919/date.2019.8715174 · Cited by 38
“[2], [3] also provide partitioning strategies to map DNNs on mobile devices.”
BackgroundDistributed artificial intelligence: Taxonomy, review, framework, and reference architecture
Janbi, N. F., Katib, I. A., & Mehmood, R. (2023) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2023.200231 · Cited by 44
DISNET: Distributed Micro-Split Deep Learning in Heterogeneous Dynamic IoT
Samikwa, E., Maio, A. D., & Braun, T. I. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3313514 · Cited by 36
“MoDNN [23] proposed input data partitioning scheme for convolutional and fully connected layers.”
BackgroundMethodInfluentialSelf-aware distributed deep learning framework for heterogeneous IoT edge devices
Jin, Y., Cai, J., Xu, J., Huan, Y., Yan, Y., Huang, B., Guo, Y., Zheng, L., & Zou, Z. (2021) · Future Generation Computer Systems · doi:10.1016/j.future.2021.07.010 · Cited by 33
Adaptive Distributed Convolutional Neural Network Inference at the Network Edge with ADCNN
Zhang, S. Q., Lin, J., & Zhang, Q. (2020) · Proceedings of the International Conference on Parallel Processing · doi:10.1145/3404397.3404473 · Cited by 34
Li, N., Iosifidis, A., & Zhang, Q. (2022) · Computer Networks · doi:10.1016/j.comnet.2022.109150 · Cited by 34
Adaptive DNN Partition in Edge Computing Environments
Miao, W., Zeng, Z., Wei, L., Li, S., Jiang, C., & Zhang, Z. (2020) · IEEE International Conference on Parallel and Distributed Systems (ICPADS) · doi:10.1109/icpads51040.2020.00097 · Cited by 34
“The authors in [15] used the idea of Map-Reduce, and they used several WiFi-connected devices to construct a small cluster.”
MethodNa, J., Zhang, H., Lian, J., & Zhang, B. (2022) · Applied Sciences · doi:10.3390/app122010619 · Cited by 30
“, partitioning the inputs [24,25], weights [26], and layers [18,19], as well as hybrid strategies [17,20,27–30].”
BackgroundKilcioglu, E., Mirghasemi, H., Stupia, I., & Vandendorpe, L. (2021) · IEEE Access · doi:10.1109/access.2021.3084689 · Cited by 27
“Among each layer-pack, convolutional layers dominate the overall energy consumption for CNNs [41], [42], [44].”
BackgroundMethodPartitioning of CNN Models for Execution on Fog Devices
Dey, S., Mukherjee, A., Pal, A., & Balamuralidhar, P. (2018) · doi:10.1145/3277893.3277899 · Cited by 32
“In future we would liketo experimentwith WifiDirect forhighercommunication speed, as done in [20].”
MethodInfluentialGenAI at the Edge: Comprehensive Survey on Empowering Edge DevicesPreprint
Navardi, M., Aalishah, R., Fu, Y., Lin, Y., Li, H., Chen, Y., & Mohsenin, T. (2025) · Qeios · doi:10.1609/aaaiss.v5i1.35586 · Cited by 39
Self-citationContext-aware Adaptive Surgery: A Fast and Effective Framework for Adaptative Model Partition
Wang, H., Guo, B., Liu, J., Liu, S., Wu, Y., & Yu, Z. (2021) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3478073 · Cited by 30
“[29] design different task assignment algorithms for different layers to minimize the interaction latency between devices, achieving a better acceleration effect.”
BackgroundMethodMemory-Reduced Network Stacking for Edge-Level CNN Architecture With Structured Weight Pruning
Moon, S., Byun, Y., Park, J., Lee, S., & Lee, Y. (2019) · IEEE Journal on Emerging and Selected Topics in Circuits and Systems · doi:10.1109/jetcas.2019.2952137 · Cited by 27
“to directly apply such CNN algorithms to edge-level devices, which are regarded as the next big market for the realization of on-device neural network processing [11].”
MethodEdgeCI: Distributed Workload Assignment and Model Partitioning for CNN Inference on Edge Clusters
Chen, Y., Luo, T., Fang, W. W., & Xiong, N. N. (2024) · ACM Transactions on Internet Technology · doi:10.1145/3656041 · Cited by 27
“Finally, we compare our proposed EdgeCI, which consists of AWAS and DPFL with the following five approaches: (1) MoDNN [24]; (2) AOFL [25]; (3) EdgeDI without streamlined transformation of CNN model [27]; (4) CoEdge [33]; (5) DistrEdge [34].”
BackgroundMethodInfluentialSelf-citationUncertainty-aware Task Offloading for Multi-vehicle Perception Fusion over Vehicular Edge Computing
Ku, Y.-J., Baidya, S., & Dey, S. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3284369 · Cited by 29
“Therefore, in this work, we propose a real-time predictive two-step based algorithm, FOrecasting and unceRtainty-aware Multi-source task Offloading and Scheduling Algorithm (FOR-MOSA) to solve problem (14).”
BackgroundMethodFograph: Enabling Real-Time Deep Graph Inference with Fog Computing
Zeng, L., Huang, P., Luo, K., Zhang, X., Zhou, Z., & Chen, X. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3511982 · Cited by 24
Self-citationHorizontally Distributed Inference of Deep Neural Networks for AI-Enabled IoT
Rodríguez, I., Campos, C., & Fdez‐Riverola, F. (2023) · Sensors · doi:10.3390/s23041911 · Cited by 28
“A third or less of the co-authored papers do not address issues such as accuracy [71,74,75,77,83] (fundamental in user-centric intelligent systems), memory footprint [71,73,74,81,87,89], and communication overhead [70,71,73,82,84,90,91] (prevalent as major challenges in the smart IoT research reviewed), or energy cons…”
InfluentialInfer-EDGE: Dynamic DNN Inference Optimization in Just-in-Time Edge-AI Implementations
Mounesan, M., Zhang, X., & Debroy, S. (2025) · IEEE/IFIP Network Operations and Management Symposium (NOMS) · doi:10.1109/noms57970.2025.11073623 · Cited by 32
DLSTM: Distributed Long Short-Term Memory Neural Networks for the Internet of Things
Wen, G., Qin, J., Fu, X., & Yu, W. (2021) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2021.3054244 · Cited by 21
Efficient Computer Vision on Edge Devices with Pipeline-Parallel Hierarchical Neural Networks
Goel, A., Tung, C., Hu, X., Thiruvathukal, G. K., Davis, J. C., & Lu, Y.-H. (2022) · Asia and South Pacific Design Automation Conference (ASP-DAC) · doi:10.1109/asp-dac52403.2022.9712574 · Cited by 24
“MoDNN [7] divides each activation map into overlapping grid cells.”
BackgroundMethodInfluentialEdgeAl: A Vision for Deep Learning in the IoT Era
Bhardwaj, K., Suda, N., & Mărculescu, R. (2019) · IEEE Design and Test · doi:10.1109/mdat.2019.2952350 · Cited by 20
DistPrivacy: Privacy-Aware Distributed Deep Neural Networks in IoT surveillance systems
Baccour, E., Erbad, A., Mohamed, A. M. S., Hamdi, M., & Guizani, M. (2020) · Globecom · doi:10.1109/globecom42002.2020.9322470 · Cited by 25
“In this way, the entire or most of the inference can be done at the proximity of data source [8], [9].”
BackgroundEdgeLD: Locally Distributed Deep Learning Inference on Edge Device Clusters
Feng, X., Fang, W. W., Xu, W., Wang, Q., Ma, X., & Ding, Y. (2020) · IEEE International Conference on High Performance Computing and Communications; IEEE International Conference on Smart City; IEEE International Conference on Data Science and Systems (HPCC/SmartCity/DSS) · doi:10.1109/hpcc-smartcity-dss50907.2020.00078 · Cited by 27
“MoDNN [3] is the first work designed to orchestrate DNN inference among multiple mobile devices in the same WLAN.”
BackgroundMethodResultInfluentialCooperative Distributed Deep Neural Network Deployment with Edge Computing
Yang, C.-Y., Kuo, J.-J., Sheu, J., & Zheng, K.-J. (2021) · International Conference on Communications · doi:10.1109/icc42927.2021.9500668 · Cited by 26
“MLS is compared with the layer-by-layer partition and slicing (SLS) algorithm based on [6], [7].”
MethodPipeline Parallelism for Inference on Heterogeneous Edge ComputingPreprint
Hu, Y., Imes, C., Zhao, X., Kundu, S., Beerel, P. A., Crago, S. P., & Walters, J. P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.14895 · Cited by 36
“Considering the limited bandwidth and uncertain delay between the edge and the cloud, MoDNN [Mao et al. , 2017] employs a MapReduce-like distributed inference paradigm and only utilizes idle mobile devices to execute CNN models.”
MethodPICO: Pipeline Inference Framework for Versatile CNNs on Diverse Mobile Devices
Yang, X., Xu, Z., Qi, Q., Wang, J., Sun, H., Liao, J., & Guo, S. (2023) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2023.3265111 · Cited by 19
“All these works [4], [22], [31], [33] require devices to communicate with each other for every layer.”
BackgroundMethodInfluentialA power-efficient and high performance FPGA accelerator for convolutional neural networks
Gong, L., Wang, C., Li, X., Chen, H., & Zhou, X. (2017) · International Conference on Hardware/Software Codesign and System Synthesis · doi:10.1145/3125502.3125534 · Cited by 20
“Along with advances in semiconductor technology as well as the evolution of CNN models, the attention of CNN in embedded and mobile devices has been gradually increased [1].”
Guo, C., Zhengqing, L., & Jie, S. (2023) · Journal of Cloud Computing Advances Systems and Applications · doi:10.1186/s13677-023-00404-y · Cited by 21
“MoDNN [13] is a locally distributed mobile computing system for DNN applications, which partition trained DNN model onto several mobile devices to accelerate DNN computations by alleviating device-level computing cost and memory usage.”
MethodMicroDeep: In-network Deep Learning by Micro-Sensor Coordination for Pervasive Computing
Fukushima, Y., Miura, D., Hamatani, T., Yamaguchi, H., & Higashino, T. (2018) · IEEE International Conference on Smart Computing · doi:10.1109/smartcomp.2018.00087 · Cited by 17
“MoDNN [7] shows the acceleration of CNN computation by partitioning pre-trained CNN models onto several mobile devices.”
MethodRL-DistPrivacy: Privacy-Aware Distributed Deep Inference for Low Latency IoT Systems
Baccour, E., Erbad, A., Mohamed, A. M. S., Hamdi, M., & Guizani, M. (2022) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2022.3165472 · Cited by 19
Memory optimization at Edge for Distributed Convolution Neural Network
Naveen, S., & Kounte, M. R. (2022) · Transactions on Emerging Telecommunications Technologies · doi:10.1002/ett.4648 · Cited by 23
Huang, Y., Qiao, X., Ren, P., Liu, L., Pu, C., & Chen, J. (2019) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs.2019.00149 · Cited by 17
“MoDNN [17] proposes two partition schemes to minimize non-parallel data delivery latency and accelerate DNN inference by alleviating device-level computing cost and memory usage.”
MethodEdge Intelligence: Challenges and Opportunities
Hu, H., & Jiang, C. (2020) · International Conference on Computer, Information and Telecommunication Systems · doi:10.1109/cits49457.2020.9232575 · Cited by 17
“Grid partition maps each computing task in the DL to a processing unit, such as MoDNN [17] and DeepThings [18].”
MethodDynamic DNN Decomposition for Lossless Synergistic Inference
Zhang, B., Tian, X., Zhang, H., Li, T., Zhu, S., & Gu, J. (2021) · IEEE International Conference on Distributed Computing Systems Workshops (ICDCSW) · doi:10.1109/icdcsw53096.2021.00010 · Cited by 22
Artificial intelligence in edge devices
Singh, A., Saini, K., Nagar, V., Aseri, V., Sankhla, M. S., Pandit, P. P., & Chopade, R. L. (2022) · Advances in computers · doi:10.1016/bs.adcom.2022.02.013 · Cited by 22
Metadata
- Authors
- Mao, J., Chen, X., Nixon, K. W., Krieger, C., & Chen, Y.
- Affiliations
- University of Pittsburgh (US)
- George Mason University (US)
- University of Maryland, College Park (US)
- Venue
- Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017
- Publisher
- IEEE
- Published
- 2017-03
- Pages
- 1396-1401
- OpenAlex
- W2612193523
- Cited by
- 331
- References
- 20
- Citations by year
- 2026: 21 · 2025: 35 · 2024: 47 · 2023: 62 · 2022: 43 · 2021: 51 · 2020: 38 · 2019: 17 · 2018: 11 · 2017: 6
- FWCI
- 15.69
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Design, Automation, and Test in Europe (publishedVersion)
- Funders
- National Science Foundation
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- deep neural network, mobile computing, distributed computing, model partitioning, resource-constrained devices, computational acceleration, parallel execution, memory reduction, worker nodes
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.