Abstract
Deep neural networks (DNNs) involving massive neural nodes grouped into different neural layers have been a promising innovation for function approximation and inference, which have been widely applied to various vertical applications such as image recognition. However, the computing burdens to train a DNN model with a limited latency may not be affordable for the user equipment (UE), which consequently motivates the concept of splitting the computations of DNN layers to not only the edge server but also the cloud platform. Despite the availability of more computing resources, computing tasks with such split computing also suffer packet transmission unreliability, latency, and significant energy consumption. A practical scheme to optimally split the computations of DNN layers to the UE, edge, and cloud is thus urgently desired. To solve this optimization, we propose a multi-tier deep reinforcement learning (DRL) scheme for the UE and edge to distributively determine the splitting points to minimize the overall training latency while meeting the constraints of overall energy consumption and image recognition accuracy. The performance evaluation results show the outstanding performance of the proposed design as compared with state-of-the-art schemes, to fully justify the practicability in the next-generation smart networks.
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig1_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig2_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig3_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig4_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig5_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11276-023-03600-5/MediaObjects/11276_2023_3600_Fig6_HTML.png)
Similar content being viewed by others
References
Li, G., Hari, S. K. S. , Sullivan, M., Tsai, T., Pattabiraman, K., Emer, J., & Keckler, S. W. (2017). Understanding error propagation in deep learning neural network (DNN) accelerators and applications. In International conference for high performance computing, networking, storage and analysis (pp. 1–12).
Chowdhary, K. (2020). Natural language processing. In Fundamentals of artificial intelligence (pp. 603–649).
5G system (5GS); study on traffic characteristics and performance requirements for AI/ML model transfer. Technical Report TS 22.874, 3GPP (2021).
Liang, T., Glossner, J., Wang, L., Shi, S., & Zhang, X. (2021). Pruning and quantization for deep neural network acceleration: A survey. Neurocomputing, 461, 370–403.
Li, E., Zeng, L., Zhou, Z., & Chen, X. (2019). Edge AI: On-demand accelerating deep neural network inference via edge computing. IEEE Transactions on Wireless Communications, 19(1), 447–457.
Heshratifar, A. E., Esmaili, A., & Pedram, M. (2019). Bottlenet: A deep learning architecture for intelligent mobile cloud computing services. IEEE/ACM ISLPED (pp. 1–6).
Li, E., Zhou, Z., & Chen, X. (2018). Edge intelligence: On-demand deep learning model co-inference with device-edge synergy. In Workshop on mobile edge communications (pp. 31–36).
Banitalebi-Dehkordi, A., Vedula, N., Pei, J., **a, F., Wang, L., & Zhang, Y. (2021). Auto-split: A general framework of collaborative edge-cloud AI. In The 27th ACM SIGKDD conference on knowledge discovery and data mining (pp. 2543–2553).
Kang, Y., Hauswald, J., Gao, C., Rovinski, A., Mudge, T., Mars, J., & Tang, L. (2017). Neurosurgeon: Collaborative intelligence between the cloud and mobile edge. ACM SIGARCH Computer Architecture News, 45(1), 615–629.
Tang, X., Chen, X., Zeng, L., Yu, S., & Chen, L. (2020). Joint multiuser DNN partitioning and computational resource allocation for collaborative edge intelligence. IEEE Internet of Things Journal, 8(12), 9511–9522.
Weissberger, A. (2022). Summary of ITR-U workshop on “IMT for 2030 and beyond” (aka“6G”).
Liu, G., Huang, Y., Li, N., Dong, J., **, J., Wang, Q., & Li, N. (2020). Vision, requirements and network architecture of 6G mobile network beyond 2030. China Communications, 17(9), 92–104.
6G vision white paper. Technical Report 1.0, MediaTek (2022).
Curnow, H. J., & Wichmann, B. A. (1976). A synthetic benchmark. The Computer Journal, 19(1), 43–49.
Acar, H., Alptekin, G. I., Gelas, J. P., & Ghodous, P. (2016). Beyond CPU: Considering memory power consumption of software. IEEE SMARTGREENS (pp. 1–8).
Chu, P. C., & Beasley, J. E. (1998). A genetic algorithm for the multidimensional knapsack problem. Journal of Heuristics, 4(1), 63–86.
Li, Z., Harman, M., & Hierons, R. M. (2007). Search algorithms for regression test case prioritization. IEEE Transactions on Software Engineering, 33(4), 225–237.
Nwankpa, C. E., Ijomah, W., Gachagan, A., & Marshall, S. (2021). Activation functions: Comparison of trends in practice and research for deep learning. In International conference on computational sciences and technology.
Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms.
Krizhevsky, A., & Hinton, G. et al. (2009). Learning multiple layers of features from tiny images.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017). Imagenet classification with deep convolutional neural networks. Communications of the ACM, 60(6), 84–90.
Strisciuglio, N., Lopez-Antequera, M., & Petkov, N. (2020). Enhanced robustness of convolutional networks with a push-pull inhibition layer. Neural Computing and Applications, 32(24), 17957–17971.
Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition.
Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications.
Guidelines for evaluation of radio interface technologies for IMT-2020. Report ITU (2017).
IEEE standard for Ethernet. (2020). Power over Ethernet over 2 pairs. IEEE: Technical Report.
Pocovi, G., Thibault, I., Kolding, T., Lauridsen, M., Canolli, R., Edwards, N., & Lister, D. (2019). On the suitability of LTE air interface for reliable low-latency applications. In IEEE WCNC (pp. 1–6).
Palumbo, F., Aceto, G., Botta, A., Ciuonzo, D., Persico, V., & Pescapé, A. (2019). Characterizing cloud-to-user latency as perceived by AWS and Azure users spread over the globe. In IEEE GLOBECOM (pp. 1–6).
Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization.
Willmott, C. J., & Matsuura, K. (2005). Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Climate Research, 30(1), 79–82.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Lien, SY., Yeh, CH. & Deng, DJ. Optimum splitting computing for DNN training through next generation smart networks: a multi-tier deep reinforcement learning approach. Wireless Netw 30, 1737–1751 (2024). https://doi.org/10.1007/s11276-023-03600-5
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11276-023-03600-5