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
Items 101–200 of 327
Edge-MultiAI: Multi-Tenancy of Latency-Sensitive Deep Learning Applications on Edge
Zobaed, S., Mokhtari, A., Champati, J. P., Kourouma, M., & Salehi, M. A. (2022) · 2022 IEEE/ACM 15th International Conference on Utility and Cloud Computing (UCC) · doi:10.1109/ucc56403.2022.00012 · Cited by 15
“proposed a mobile computing framework, MoDNN, to execute DL applications simultaneously on resource-constrained devices [15].”
MethodServing Graph Neural Networks With Distributed Fog Servers for Smart IoT Services
Zeng, L., Chen, X., Huang, P., Luo, K., Zhang, X., & Zhou, Z. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3293052 · Cited by 21
“To achieve efficient intelligent services close to the end, a line of works [77], [78], [79], [80], [81], [82] have explored collaborative execution with the assist of fog computing.”
BackgroundSelf-citationHenna, S., & Davy, A. (2020) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2020.2978846 · Cited by 17
“There is some recent work on the deployment of deep neural networks on mobile devices within a WLAN [34].”
BackgroundDeep learning for internet of things in fog computing: Survey and Open Issues
Tmamna, J., Ayed, E. B., & Ayed, M. B. (2020) · International Conference on Advanced Technologies for Signal and Image Processing · doi:10.1109/atsip49331.2020.9231685 · Cited by 18
“Results Critics [46], [47], [48], [49], [50] Compression Reduce energy, memory Computational requirement still high, loss accuracy [51], [52] offloading Latency, accuracy Latency still high [53], [54] partitioning Latency, energy high communication cost”
BackgroundMethodA Distributed In-Situ CNN Inference System for IoT Applications
Du, J., Shen, M., & Du, Y. (2020) · IEEE International Conference on Computer Design: VLSI in Computers and Processors (ICCD) · doi:10.1109/iccd50377.2020.00055 · Cited by 16
“D grid partitioning method from MoDNN [15] and the 2-”
BackgroundMethodInfluentialAMVP: Adaptive CNN-based Multitask Video Processing on Mobile Stream Processing Platforms
Chao, M., Stoleru, R., Jin, L., Yao, S., Maurice, M., & Blalock, R. (2020) · IEEE/ACM Symposium on Edge Computing (SEC) · doi:10.1109/sec50012.2020.00015 · Cited by 19
“Modnn [14] and Mednn [15] utilize specific partition schemes to partition CNN models onto several mobile devices to alleviate device-level computing cost and memory footprint.”
FeSHI: Feature Map-Based Stealthy Hardware Intrinsic Attack
Odetola, T. A., Khalid, F., Mohammed, H., Sandefur, T., & Hasan, S. R. (2021) · IEEE Access · doi:10.1109/access.2021.3104520 · Cited by 15
Ding, Y., Fang, W. W., Liu, M., Wang, M., Cheng, Y., & Xiong, N. N. (2022) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2022.11.008 · Cited by 16
Map-and-Conquer: Energy-Efficient Mapping of Dynamic Neural Nets onto Heterogeneous MPSoCs
Bouzidi, H., Odema, M., Ouarnoughi, H., Niar, S., & Faruque, M. A. A. (2023) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac56929.2023.10247722 · Cited by 27
The Case for Adaptive Deep Neural Networks in Edge Computing
McNamee, F., Dustdar, S., Kilpatrick, P., Shi, W., Spence, I. T. A., & Varghese, B. (2021) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud53861.2021.00017 · Cited by 16
“These methods include, estimation-based techniques [6], [13], [14], structural modification-based techniques [15], [16], and measurement-based techniques [17].”
MethodJoint compressing and partitioning of CNNs for fast edge-cloud collaborative intelligence for IoT
Zhang, W., Wang, N., Li, L., & Wei, T. (2022) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2022.102461 · Cited by 15
An Autonomous Parallelization of Transformer Model Inference on Heterogeneous Edge Devices
Lee, J., Bahk, I., Kim, H., Jeong, S., Lee, S., & Min, D. (2024) · International Conference on Supercomputing · doi:10.1145/3650200.3656628 · Cited by 20
All-Sky Autonomous Computing in UAV Swarm
Sun, H., Qu, Y., Dong, C., Dai, H., Li, Z., Zhang, L., Wu, Q., & Guo, S. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3427420 · Cited by 15
“Like DeepSlicing, MoDNN [42] considered each convolution layer combined with the subsequent non-overlapping layer structure as a subtask.”
MethodPeng, S., Shen, Z., Zheng, Q., Hou, X., Jiang, D., Yuan, J., & Jin, J. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3585287 · Cited by 21
Towards Efficient Inference: Adaptively Cooperate in Heterogeneous IoT Edge Cluster
Yang, X., Qi, Q., Wang, J., Guo, S., & Liao, J. (2021) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs51616.2021.00011 · Cited by 16
Split-Et-Impera: A Framework for the Design of Distributed Deep Learning Applications
Capogrosso, L., Cunico, F., Lora, M., Cristani, M., Fummi, F., & Quaglia, D. (2023) · IEEE International Symposium on Design and Diagnostics of Electronic Circuits & Systems (DDECS) · doi:10.1109/ddecs57882.2023.10139711 · Cited by 13
“, BODP, MSCC), to enable parallel computation of DNN on mobile platforms is presented in [15].”
BackgroundContext Recognition of Humans and Objects by Distributed Zero-Energy IoT Devices
Teruo Higashino, Y., Uchiyama, A., Saruwatari, S., Yamaguchi, H., & Watanabe, T. (2019) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs.2019.00177 · Cited by 13
Adeyemo, A. A., Sanderson, J. J., Odetola, T. A., Khalid, F., & Hasan, S. R. (2023) · IEEE Access · doi:10.1109/access.2023.3241096 · Cited by 12
Optimization framework for splitting DNN inference jobs over computing networks
Jung, S., & Lee, H.-W. (2023) · Computer Networks · doi:10.1016/j.comnet.2023.109814 · Cited by 15
“In [15], [16], even a layer in the fully connected part is partitioned into multiple groups and distributed to computing nodes.”
BackgroundCai, S., Wang, D., Wang, H., Lyu, Y., Xu, G., Zheng, X., & Vasilakos, A. V. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118419 · Cited by 13
“Moreover, MoDNN [24] and DeepThings [25] both adopt fine-grained partitioning to distribute DNN executions on lightweight end devices based on the computation capabilities and the memory usage of the end devices during run-time.”
BackgroundPang, B., Liu, S., Wang, H., Guo, B., Wang, Y., Wang, H., Sheng, Z., Wang, Z., & Yu, Z. (2023) · ACM Transactions on Sensor Networks · doi:10.1145/3630098 · Cited by 11
“Existing efforts explore partitioning a DNN at several points and computing partial modules on different edge devices in series [21, 26] (partitioned between DNN layers) or parallel [34, 62] (partitioned within layers).”
BackgroundXu, Z., Qin, Z., Yu, F., Liu, C., & Chen, X. (2018) · International Symposium on Low Power Electronics and Design (ISLPED) · doi:10.1145/3218603.3218652 · Cited by 11
Self-citationEnergy-efficient neural networks with near-threshold processors and hardware accelerators
Nunez-Yanez, J. L., & Howard, N. (2021) · Journal of Systems Architecture · doi:10.1016/j.sysarc.2021.102062 · Cited by 13
Hardening Hardware Accelerartor Based CNN Inference Phase Against Adversarial Noises
Odetola, T. A., Adeyemo, A. A., & Hasan, S. R. (2022) · IEEE International Symposium on Hardware Oriented Security and Trust (HOST) · doi:10.1109/host54066.2022.9840017 · Cited by 13
ResMap: Exploiting Sparse Residual Feature Map for Accelerating Cross-Edge Video Analytics
Chen, N., Zhang, S., Zhang, S., Yan, Y., Chen, Y., & Lu, S. (2023) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom53939.2023.10228990 · Cited by 12
“To mitigate the insufficient computation and memory at the edges but accelerate the CNN inference, a variety of approaches have been studies, such as model compression [6]– [8], model early-exit [9]–[11], model partitioning [12]–[15], [20]–[23], data partitioning [25]–[29], and specific hardware or tools [30], [31].”
MethodSelf-citationSynergy: Towards On-Body AI via Tiny AI Accelerator Collaboration on Wearables
Gong, T., Jang, S. Y., Acer, U. G., Kawsar, F., & Min, C. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3564314 · Cited by 20
“Conversely, except for IndE2E, the throughput gains for all other baselines are not notable, even with additional devices.”
Low-Complexity Dynamic Channel Scaling of Noise-Resilient CNN for Intelligent Edge Devices
Byun, Y., Ha, M., Kim, J., Lee, S., & Lee, Y. (2019) · Design, Automation, and Test in Europe · doi:10.23919/date.2019.8715280 · Cited by 12
“Following to the high-performance computing servers, dominating the ML market with massive-parallel operations [2], the edge devices using the raw inputs are now expected to have the intelligence by including CNN inference engines [3].”
MethodNazir, A., Mir, R. N., & Qureshi, S. (2020) · International Journal of Intelligent Computing and Cybernetics · doi:10.1108/ijicc-04-2020-0038 · Cited by 13
Huang, Q., Yang, H., Zeng, E., & Chen, Y. (2024) · IEEE/ACM Conference on Connected Health: Applications, Systems and Engineering Technologies · doi:10.1109/chase60773.2024.00020 · Cited by 14
“To enable wider and more efficient deployment of deep-learning-based methods, model compression and efficient computation techniques like pruning [48], [49], quantization [50], parallel computation [51], and neural architecture search [52] on vital signal processing tasks are worth investigating.”
BackgroundSelf-citationHiDP:Hierarchical DNN Partitioning for Distributed Inference on Heterogeneous Edge Platforms
Taufique, Z., Vyas, A., Miele, A., Liljeberg, P., & Kanduri, A. (2025) · Design, Automation, and Test in Europe · doi:10.23919/date64628.2025.10992692 · Cited by 20
“Existing distributed edge inference strategies partition the inference workload into blocks by splitting the layers of a DNN model [6]–[12] and/or the input data of the DNN inference request [3], [4], [13]–[16].”
InfluentialOffloading Methodologies for Energy consumption in Mobile Edge Computing
Saranya, G., & Sasikala, E. (2021) · International Conference on Smart Electronics and Communication (ICOSEC) · doi:10.1109/icosec51865.2021.9591883 · Cited by 10
“MoDNN [52] and Deep Things [50], for example, To spread DNN executions on lightweight end devices like Android smart phones and Raspberry Pies, utilizing fine-grained partitioning.”
MethodCooperative Convolutional Neural Network Deployment over Mobile Networks
Hsu, C.-C., Yang, C.-K., Kuo, J.-J., Chen, W.-T., & Sheu, J. (2020) · International Conference on Communications · doi:10.1109/icc40277.2020.9149094 · Cited by 12
“We compare THREAD-SA and THREAD with MoDNN [8] with single antenna and multiple antennas (i.”
BackgroundMethodInfluentialRL-PDNN: Reinforcement Learning for Privacy-Aware Distributed Neural Networks in IoT Systems
Baccour, E., Erbad, A., Mohamed, A. M. S., Hamdi, M., & Guizani, M. (2021) · IEEE Access · doi:10.1109/access.2021.3070627 · Cited by 13
“On the other hand, two contributions marked the input-wise partitioning, namely MoDNN [4] and DeepThings [5].”
Security Analysis of Capsule Network Inference using Horizontal Collaboration
Adeyemo, A. A., Khalid, F., Odetola, T. A., & Hasan, S. R. (2021) · IEEE International Midwest Symposium on Circuits and Systems (MWSCAS) · doi:10.1109/mwscas47672.2021.9531833 · Cited by 10
“To accommodate the high computation and memory requirement for the deployment of the CapsNet on hardware accelerators like FPGAs, horizontal collaborative inference [13], [2], [14] could be employed.”
MethodCommunication-Aware DNN Pruning
Jian, T., Roy, D., Salehi, B., Soltani, N., Chowdhury, K. R., & Ioannidis, S. (2023) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom53939.2023.10229043 · Cited by 13
“…without considering their physical network topology: literature on partitioning a DNN across edge devices typically assumes all-to-all connections [10]–[13], or no connections until the last layer [14]–[17]; the To reduce latency and accelerate computation, inference over a DNN is distributed…”
BackgroundFailure-Resilient Distributed Inference With Model Compression Over Heterogeneous Edge Devices
Wang, L., Li, L., Xu, L., Peng, X., & Fei, A. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3419831 · Cited by 13
“Mao et al. in [9] enable execution parallelism among multiple mobile devices by partitioning the neurons of each layer, requiring the transfer of overlapped parts of layer inputs during computation.”
BackgroundSelf-aware collaborative edge inference with embedded devices for IIoT
Chen, Y., Yu, Z., Jin, Y., Mwase, C., Hu, X. P., Xu, L. D., Zou, Z., & Zheng, L. (2024) · Future Generation Computer Systems · doi:10.1016/j.future.2024.107535 · Cited by 11
Self-citationEdgeBoost: Confidence boosting for resource constrained inference via selective offloading
Said, N., & Landsiedel, O. (2025) · Computer Networks · doi:10.1016/j.comnet.2025.111437 · Cited by 10
Tu, C.-H., Sun, Q., & Cheng, M.-H. (2021) · The Journal of Supercomputing · doi:10.1007/s11227-021-03795-4 · Cited by 10
“As for the data-partitioning approaches [25, 44], they emphasized the computation accelerations of a given CNN model with multiple end/edge devices, which means the approaches in this group are more like those introduced in Sect.”
BackgroundMethodLiao, Z., Zhang, X., He, S., & Tang, Q. (2023) · Journal of Network and Computer Applications · doi:10.1016/j.jnca.2023.103720 · Cited by 10
Optimizing the IoT Performance: A Case Study on Pruning a Distributed CNN
Saqib, E., Leal, I. S., Shallari, I., Jantsch, A., Krug, S., & O’Nils, M. (2023) · IEEE Sensors Applications Symposium (SAS) · doi:10.1109/sas58821.2023.10254054 · Cited by 12
Low-memory and high-performance CNN inference on distributed systems at the edge
Tang, E., & Stefanov, T. P. (2021) · IEEE/ACM International Conference on Utility and Cloud Computing Companion (UCC Companion) · doi:10.1145/3492323.3495629 · Cited by 12
“An example of a methodology utilizing this partitioning strategy to deploy a CNN model and perform inference on a distributed system at the edge can be found in [4, 10, 13].”
MethodMuslim, N., Islam, S., & Grégoire, J. (2020) · International Journal of Network Management · doi:10.1002/nem.2146 · Cited by 9
“Different methods 66,67 have been developed to partition a pretrained DNN over the mobile device and the cloud server to minimize computation time.”
MethodMachine Learning Inference on Serverless Platforms Using Model Decomposition
Gallego, A., Odyurt, U., Cheng, Y., Wang, Y., & Zhao, Z. (2023) · IEEE/ACM International Conference on Utility and Cloud Computing (UCC) · doi:10.1145/3603166.3632535 · Cited by 12
BPS: Batching, Pipelining, Surgeon of Continuous Deep Inference on Collaborative Edge Intelligence
Hou, X., Guan, Y., Choi, N., & Han, T. (2024) · IEEE Transactions on Cloud Computing · doi:10.1109/tcc.2024.3399616 · Cited by 10
ParaLoupe: Real-Time Video Analytics on Edge Cluster via Mini Model Parallelization
Wang, H., Li, Q., Kang, H., Hu, D., Ma, L., Tyson, G., Yuan, Z., & Jiang, Y. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3438155 · Cited by 10
“The idea of leveraging multiple cooperative devices for distributed inference has also been adopted by previous works via data partition [6], [7], [20] and model partition [21], [22], [23], [24].”
BackgroundMethodSmart Split: Leveraging TinyML and Split Computing for Efficient Edge AI
Bove, F., & Bedogni, L. (2024) · IEEE/ACM Symposium on Edge Computing (SEC) · doi:10.1109/sec62691.2024.00052 · Cited by 10
“The MoDNN system [16] presents a local distributed mobile computing approach for deep neural networks.”
Li, Y., Li, K., Wei, W., Zhou, J. T., & Chen, C. (2022) · ACM Transactions on Sensor Networks · doi:10.1145/3526191 · Cited by 11
“To alleviate the pressure of cloud center, the traditional partitioning approaches [30] often use row/columnand grid-based strategies for distributed CNN inference, However, they often assume the entire inputs are dynamically generated and do not consider the static characteristics of historical records.”
MethodPTEENet: Post-Trained Early-Exit Neural Networks Augmentation for Inference Cost Optimization
Lahiany, A., & Aperstein, Y. (2022) · IEEE Access · doi:10.1109/access.2022.3187002 · Cited by 15
“In the case of DNN, an edge device might compute few lower layers of the network, while transmitting the output of an intermediate layer to the server for the rest of the processing [21].”
BackgroundResearching the CNN Collaborative Inference Mechanism for Heterogeneous Edge Devices
Wang, J., Chen, C., Li, S., Wang, C., Cao, X., & Yang, L. (2024) · Sensors · doi:10.3390/s24134176 · Cited by 10
“The literature [11] designed a local distributed computing system that partitions a trained DNN (Deep Neural Network) model across multiple mobile devices to accelerate inference.”
MethodWang, X., Alassafi, M. O., Alsaadi, F. E., Xue, X., Zou, L., & Liu, Z. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3445642 · Cited by 11
“This includes sequential [12], parallel [13], and hybrid collaborative computing modes [14] based on different deep model segmentation strategies.”
BackgroundBolat, Y., Murray, I. D., Ren, Y., & Ferdosian, N. (2025) · Sensors · doi:10.3390/s25154595 · Cited by 12
Bhardwaj, K., Chen, W., & Mărculescu, R. (2020) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac18072.2020.9218628 · Cited by 9
“However, MoDNN assumes that the entire model can fit on each device, and does not consider any model compression.”
BackgroundAdaLearner: An adaptive distributed mobile learning system for neural networks
Mao, J., Qin, Z., Xu, Z., Nixon, K. W., Chen, X., Li, H. H., & Chen, Y. (2017) · International Conference on Computer Aided Design · doi:10.1109/iccad.2017.8203791 · Cited by 8
Self-citationTowards Enabling Dynamic Convolution Neural Network Inference for Edge Intelligence
Adeyemo, A. A., Sandefur, T., Odetola, T. A., & Hasan, S. R. (2022) · IEEE International Symposium on Circuits and Systems proceedings · doi:10.1109/iscas48785.2022.9937326 · Cited by 11
“on mobile platforms while balancing worker node workloads and reducing data delivery time was introduced in [11].”
BackgroundVelasco-Montero, D., Goossens, B., Fernández‐Berni, J., Rodriguez-Vazquez, A., & Philips, W. R. (2023) · IEEE Access · doi:10.1109/access.2023.3264828 · Cited by 11
“Multi-node edge-edge solutions can also optimize network execution by applying layer fusion, data parallelism, or network partition over clusters of edge devices [18], [19], [30], [31], [32], [33].”
BackgroundPower- and Endurance-Aware Neural Network Training in NVM-Based Platforms
Meng, F., Xue, Y., & Yang, C. (2018) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2018.2858360 · Cited by 8
“In [21], a mobile platform-based deep NN (DNN)”
BackgroundHierarchical deployment of deep neural networks based on fog computing inferred acceleration model
Jiang, W., & Lv, S. (2021) · Cluster Computing · doi:10.1007/s10586-021-03298-0 · Cited by 10
“…and Huffman coding, which reduces the storage requirements of the DNN model without affecting the accuracy of the DNN model; In the reference [9], aiming at the problem of excessive memory space and resources required for convolutional neural network calculations in embedded devices with…”
MethodEnergy-Efficient Mapping for a Network of DNN Models at the Edge
Ghasemi, M., Heidari, S., Kim, Y. G., Lamb, A., Wu, C.-J., & Vrudhula, S. B. K. (2021) · IEEE International Conference on Smart Computing · doi:10.1109/smartcomp52413.2021.00024 · Cited by 8
A Locally Distributed Mobile Computing Framework for DNN based Android Applications
Zhang, J., Chen, S., Liu, B., Ma, Y., & Chen, X. (2018) · Proceedings of the Tenth Asia-Pacific Symposium on Internetware · doi:10.1145/3275219.3275236 · Cited by 7
“Due to communication overhead, the total execution time does not decrease linearlywith the increase ofworking nodes [12].”
BackgroundSelf-citationDistributed Computation of DNN via DRL With Spatiotemporal State Embedding
Kim, S., Jung, S., & Lee, H.-W. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3336695 · Cited by 9
“For example, MoDNN [10] and DeepThings [9] employ fine-grained partitioning to harness lightweight devices like Raspberry Pis that cannot carry out heavy computations on its own resource.”
BackgroundA Survey of AI Inference Technologies for On-Device Systems
Wang, W., Li, K., Ji, B., Liu, X., Yu, J., & Wu, Q. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3613691 · Cited by 16
“Compared to MoDNN [114], DISNET improves inference speed by 5 .”
InfluentialJoint Architecture Design and Workload Partitioning for DNN Inference on Industrial IoT Clusters
Fang, W. W., Xu, W., Yu, C., & Xiong, N. N. (2022) · ACM Transactions on Internet Technology · doi:10.1145/3551638 · Cited by 11
“MoDNN [26] is the pioneer work for orchestrating DNN inference among a cluster of connected edge devices.”
BackgroundMethodInfluentialModel and system robustness in distributed CNN inference at the edge
Guo, X., Jiang, Q., Pimentel, A. D., & Stefanov, T. P. (2024) · Integration · doi:10.1016/j.vlsi.2024.102299 · Cited by 7
“In such a planar CNN distribution paradigm , model partitioning [7, 8] and data partitioning[9, 10] methods are typically applied to alleviate the discrepancy between the constrained resources of edge devices and the huge requirements of deploying large CNN models.”
MethodEC5: Edge–cloud collaborative computing framework with compressive communication
Tan, J., Liu, F., Wang, B., Wu, Q., & Chen, C. L. P. (2025) · Future Generation Computer Systems · doi:10.1016/j.future.2025.107715 · Cited by 9
Jigsaw: Taming BEV-centric Perception on Dual-SoC for Autonomous Driving
Sun, L., Li, C., Hou, X., Huang, T., Xu, C., Wang, X., Bao, G., Sun, B., Rui, S., & Guo, M. (2024) · IEEE Real-Time Systems Symposium (RTSS) · doi:10.1109/rtss62706.2024.00032 · Cited by 9
“It is applied to speedup LLM training/inference in datacenter [83], [84] or to enable collaborative edge CNN inference on embedded single-board computers 11 290 (SBC) [85]–[87].”
Genetic Algorithm-Based Online-Partitioning BranchyNet for Accelerating Edge Inference
Na, J., Zhang, H., Lian, J., & Zhang, B. (2023) · Sensors · doi:10.3390/s23031500 · Cited by 9
“Firstly, considering the requirements for storing large inputs or weights, inputs partitioning [23,24] and weights partitioning [25] approaches are proposed.”
BackgroundRodríguez, I., Campos, C., & Fdez‐Riverola, F. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126835 · Cited by 7
FixCaffe: Training CNN with Low Precision Arithmetic Operations by Fixed Point Caffe
Guo, S., Wang, L., Chen, B., Dou, Q., Tang, Y., & Li, Z. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-67952-5_4 · Cited by 10
ScissionLite: Accelerating Distributed Deep Neural Networks Using Transfer LayerPreprint
Ahn, H., Lee, M., Hong, C.-H., & Varghese, B. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.02019 · Cited by 7
“Although this approach is effective, they can be intrusive methods requiring modification of the underlying libraries [21], [22].”
A collaborative inference strategy for medical image diagnosis in mobile edge computing environment
Zhang, S., Cui, Y., Xu, D., & Lin, Y. (2025) · PeerJ Computer Science · doi:10.7717/peerj-cs.2708 · Cited by 8
“Mao et al. (2017) proposed MoDNN, a locally distributed mobile computing system for DNN applications, which partitions an already trained DNN model across multiple mobile devices to accelerate DNN computation by reducing device level computational cost and memory consumption.”
DISSEC: A distributed deep neural network inference scheduling strategy for edge clusters
Li, Q., Huang, L., Tong, Z., Du, T., Zhang, J., & Wang, S. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.05.084 · Cited by 8
Distributed Deep Learning Inference Acceleration using Seamless Collaboration in Edge Computing
Li, N., Iosifidis, A., & Zhang, Q. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9839083 · Cited by 6
“To allow multiple ESs to process sub-tasks simultaneously, layer-wise parallelization was used in the previous works [1], [8].”
BackgroundMethodInfluentialNetwork-level Design Space Exploration of Resource-constrained Networks-of-Systems
Zhao, Z., Barijough, K. M., & Gerstlauer, A. (2020) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3387918 · Cited by 6
“Recently, with increasing application of deep learning methods in such scenarios, different approaches have been proposed to adopt distributed Deep/Convolutional Neural Network (DNN/CNN) frameworks on mobile, edge, or IoT devices [20, 23, 40, 47].”
BackgroundEdgeWise: Energy-efficient CNN Computation on Edge Devices under Stochastic Communication Delays
Ghasemi, M., Rakhmatov, D. N., Wu, C.-J., & Vrudhula, S. B. K. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3530908 · Cited by 8
“MoDNN [31] and DeepThings [43] are instances of frameworks that distribute the computation of input data among a set of mobile devices connected using a wireless connection.”
LoADPart: Load-Aware Dynamic Partition of Deep Neural Networks for Edge Offloading
Liu, H., Zheng, W., Li, L., & Guo, M. (2022) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs54860.2022.00053 · Cited by 11
“Orthogonal to our work, MoDNN [21], DeepThings [22], and CoEdge [23] focus on partitioning computation inside each layer of a DNN and offloading the partitions to local worker nodes.”
MethodDistributed Inference Models and Algorithms for Heterogeneous Edge Systems Using Deep Learning
Yuan, Q., & Li, Z. (2025) · Applied Sciences · doi:10.3390/app15031097 · Cited by 7
DNN Partitioning for Cooperative Inference in Edge Intelligence: Modeling, Solutions, Toolchains
Hao, Y., Ding, N., Xia, W., Ge, H., & Xu, L. (2025) · ACM Computing Surveys · doi:10.1145/3786145 · Cited by 12
“As discussed in Section 2.1, the inherent parallelism of matrix computations enables fine-grained inference parallelization within DNNs. Sub-layer-based partitioning methods arise because dividing each DNN layer into smaller computational units facilitates more efficient parallel execution.”
InfluentialIntelligence Inference on IoT Devices
Zhang, Q., Li, Y., Zhang, D., Murturi, I., Pujol, V. C., Dustdar, S., & Wang, S. (2023) · Springer eBooks · doi:10.1007/978-3-031-50514-0_9 · Cited by 7
Zheng, Y., Cui, L., Tso, F. P., Li, Z., & Jia, W. (2024) · Computer Networks · doi:10.1016/j.comnet.2024.110607 · Cited by 5
Split Edge-Cloud Neural Networks for Better Adversarial Robustness
Douch, S., Abid, M. R., Zine-Dine, K., Bouzidi, D., & Benhaddou, D. (2024) · IEEE Access · doi:10.1109/access.2024.3487435 · Cited by 6
“Edge Intelligence employs a variety of techniques that allow AI models to be hosted at the edge, including quantization [7], pruning [8], sparsification [9], knowledge distillation [10], and distributed inference [11], [12], [13].”
BackgroundEnergy-Efficient, Delay-Constrained Edge Computing of a Network of DNNs
Ghasemi, M., Heidari, S., Kim, Y. G., Wu, C.-J., & Vrudhula, S. B. K. (2024) · IEEE Transactions on Computers · doi:10.1109/tc.2024.3500368 · Cited by 6
“MoDNN [31] and DeepThings [44] target the offloading problem for a single DNN using another approach.”
SHEATH: Defending Horizontal Collaboration for Distributed CNNs Against Adversarial Noise
Asif, M., Kazmi, M. K., Rahman, M. A., Hasan, S. R., & Homsi, S. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3556349 · Cited by 8
DFSNet: Dividing-fuse deep neural networks with searching strategy for distributed DNN architecture
Hou, W., Liu, L., Zhang, H., Sun, H., & Zheng, N. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.08.144 · Cited by 6
CAMDNN: Content-Aware Mapping of a Network of Deep Neural Networks on Edge MPSoCs
Heidari, S., Ghasemi, M., Kim, Y. G., Wu, C.-J., & Vrudhula, S. B. K. (2022) · IEEE Transactions on Computers · doi:10.1109/tc.2022.3207137 · Cited by 10
“MoDNN [28] and DeepThings [44] were designed to execute DNNs on resource-constrained mobile devices.”
MethodMJOA-MU: End-to-edge collaborative computation for DNN inference based on model uploading
Yang, H., Sun, S., Liu, M., Zhang, Q., & Wang, Y. (2023) · Computer Networks · doi:10.1016/j.comnet.2023.109801 · Cited by 5
Adeyemo, A. A., & Hasan, S. R. (2023) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/3583781.3590299 · Cited by 5
“On the other hand, in MP-HC, the trained model is partitioned across collaborating devices, and each device performs partial inference computations before passing the resulting FMs to the next device as shown in Fig.1b [5].”
BackgroundMethodRobustDiCE: Robust and Distributed CNN Inference at the Edge
Guo, X., Jiang, Q., Pimentel, A. D., & Stefanov, T. P. (2024) · Asia and South Pacific Design Automation Conference (ASP-DAC) · doi:10.1109/asp-dac58780.2024.10473970 · Cited by 6
“In such a horizontal CNN distribution paradigm , model partitioning [7], [8] and data partitioning [9], [10] methods are typically applied to alleviate the discrepancy between the constrained resources of edge devices and the huge requirements of deploying large CNN models.”
MethodAdaptive Workload Distribution for Accuracy-aware DNN Inference on Collaborative Edge Platforms
Taufique, Z., Miele, A., Liljeberg, P., & Kanduri, A. (2024) · Asia and South Pacific Design Automation Conference (ASP-DAC) · doi:10.1109/asp-dac58780.2024.10473987 · Cited by 7
“Most research works [10], [11] evenly partition the workload and are suitable for homogeneous edge clusters.”
BackgroundResultInfluentialDeep Neural Networks on Mobile Healthcare Applications: Practical Recommendations
Benedetto, J. I., Sanabria, P., Neyem, A., Navón, J., Poellabauer, C., & Xia, B. (2018) · International Conference on Ubiquitous Computing and Ambient Intelligence (UCAmI) · doi:10.3390/proceedings2190550 · Cited by 7
“Following a similar idea, MoDNN [21] focuses on local partitioning of already trained DNNs across multiple devices to obtain speedups.”
MethodJoint Optimization of Data Transfer and Co-Execution for DNN in Edge Computing
Fu, Z., Zhou, Y., Wu, C., & Zhang, Y. (2021) · International Conference on Communications · doi:10.1109/icc42927.2021.9500513 · Cited by 6
“Thus, redundant calculations are introduced to generate overlapped data for the next layer.”
BackgroundMethodEdge Computing for Artificial Intelligence
Wang, X., Han, Y., Leung, V. C. M., Niyato, D. T., Yan, X., & Chen, X. (2020) · Springer eBooks · doi:10.1007/978-981-15-6186-3_7 · Cited by 5
Self-citationScission: Context-aware and Performance-driven Edge-based Distributed Deep Neural Networks.Preprint
Lockhart, L., Harvey, P., Imai, P., Willis, P., & Varghese, B. (2020) · arXiv (Cornell University) · Cited by 6
CLAN: Continuous Learning using Asynchronous Neuroevolution on Commodity Edge Devices
Mannan, P., Samajdar, A., & Krishna, T. (2020) · IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS) · doi:10.1109/ispass48437.2020.00019 · Cited by 4
“some work done in distributing DNNs on the edge devices [46], [47] but such techniques are dependent on DNN primitives such as convolution layers and cannot be extended directly to the computation graph of NeuroEvolutionary algorithms.”
BackgroundMobile-YOLO: A Lightweight YOLO for Road Crack Detection on Mobile Devices
Yu, A., Editorial Office, Gao, Y., Xiong, Y., Liu, W., & She, J. (2025) · Journal of Advanced Computational Intelligence and Intelligent Informatics · doi:10.20965/jaciii.2025.p1443 · Cited by 4
Gong, L., Wang, C., Li, X., Chen, H., & Zhou, X. (2017) · International Conference on Hardware/Software Codesign and System Synthesis · Cited by 4
Communication Failure Resilient Distributed Neural Network for Edge Devices
Jeong, J., Park, J., & Yang, H. (2021) · Electronics · doi:10.3390/electronics10141614 · Cited by 4
“Two different distribution approaches for running a single huge NN applications on multiple devices: (a) MoDNN [11] and (b) Network of Neural Network [12].”
BackgroundMethodInfluentialSDPMP: Inference Acceleration of CNN Models in Heterogeneous Edge Environment
Han, B., Dai, P., Li, K., Zhao, K., & Lei, X. (2024) · World Conference on Computing and Communication Technologies (WCCCT) · doi:10.1109/wccct60665.2024.10541524 · Cited by 5
“[6], [7] consider offloading tasks to closer edge devices for collaborative inference through intra-layer partitioning, but the impact of different device placement sequences on throughput due to heterogeneous computational performance and network bandwidth has not been explored.”
BackgroundSecoInfer: Secure DNN End-Edge Collaborative Inference Framework Optimizing Privacy and Latency
Yao, Y., Hou, J., Wu, G., Cheng, Y., Yuan, M., Luo, P., Wang, Z., & Li, X. (2024) · ACM Transactions on Sensor Networks · doi:10.1145/3694972 · Cited by 6
Alghareb, F. S., & Hasan, B. T. (2025) · Computers · doi:10.3390/computers14010029 · Cited by 5
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.