Abstract
Deep learning parallel plays an important role in accelerating model training and improving prediction accuracy. In order to fully consider the authenticity of the simulation application scenario of model, the development of deep learning model is becoming more complex and deeper. However, a more complex and deeper model requires a larger amount of computation compared to common spatio-temporal model. In order to speed up the calculation speed and accuracy of the deep learning model, this work optimizes the common spatial-temporal model in deep learning from three aspects: data parallel, model parallel and gradient accumulation algorithm. Firstly, the data parallel slicing algorithm proposed in this work achieves parallel GPUs load balancing. Secondly, this work independently parallelizes the components of the deep spatio-temporal. Finally, this work proposes a gradient accumulation algorithm based on deep reinforcement learning. This work uses two data sets (GeoLife and Chengdu Taxi) to train and evaluate multiple parallel modes. The parallel mode combining data parallel and gradient accumulation algorithm is determined. The experimental effect has been greatly improved compared with the original model.
This research was supported in part by Shandong Province colleges and universities youth innovation technology plan innovation team project under Grant No. 2020KJN011, Shandong Provincial Natural Science Foundation under Grant No. ZR2020MF060, Program for Innovative Postdoctoral Talents in Shandong Province under Grant No. 40618030001, National Natural Science Foundation of China under Grant No. 61802216, and Postdoctoral Science Foundation of China under Grant No.2018M642613.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Martineau, M., Atkinson, P., McIntosh-Smith, S.: Benchmarking the NVIDIA V100 GPU and Tensor Cores. In: Mencagli, G., B. Heras, D., Cardellini, V., Casalicchio, E., Jeannot, E., Wolf, F., Salis, A., Schifanella, C., Manumachu, R.R., Ricci, L., Beccuti, M., Antonelli, L., Garcia Sanchez, J.D., Scott, S.L. (eds.) Euro-Par 2018. LNCS, vol. 11339, pp. 444–455. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-10549-5_35
Zhu, H., Li, P., Jiao, L.: Review of parallel deep neural network. Chin. J. Comput. 41(8), 1861–1881 (2018)
Lui, Q., Zhuo, J., Zhang, Z.: A survey on deep reinforcement learning. Chin. J. Comput. 41(1), 1–27 (2018)
Shu, J., Zheng, W.: Performance analysis for massive problem data parallel computing. J. Softw. 11(5), 628–633 (2000)
Alexander, S.: An architecture for parallel topic models. In: Proceedings of the VLDB Endowment, VLDB, Santa Clara, CA, vol. 3, pp. 703–710 (2010)
Wang, S.: Research on parameter-exchanging optimizing mechanism in distributed deep learning. Huazhong University of Science and Technology, pp. 1–64 (2015)
Chen, M., Yan, Z., Ye, Y.: Parallel optimization for deep learning based on HPC environment. Comput. Eng. Sci. 40(8), 133–140 (2019)
Zheng, S., Meng, Q., and Wang, T.: Asynchronous stochastic gradient descent with delay compensation. In: 34th International Conference on Machine Learning, ICML, Sydney, Australia, vol. 70, pp. 4120-4129 (2017)
Wang, L., Yang, J., Cheng, L.: Parallel optimization of chinese language model based on recurrent neural network. J. Appl. Sci. 33(3), 253–261 (2015)
Jia, Z., Zaharia, M., Aiken, A.: Beyond data and model parallelism for deep neural networks. In: 35th International Conference on Machine Learning, ICML, Stockholm, Sweden, pp. 1-15 (2018)
Lee, S., Kim, J.K., Zheng, X.: Primitives for dynamic big model parallelism, 1–22. arXiv (2014)
Shoeybi, M., Patwary, M., Puri, R.: Megatron-lm: training multi-billion parameter language models using model parallelism. arXiv, 1–15 (2019)
Ding, Y., Liu, B.: Using GPU for high-performance data parallel computing. Programmer 4(62), 97–99 (2008)
Sun, S., Chen, W., Bian, J., Liu, X., Liu, T.-Y.: Ensemble-compression: A new method for parallel training of deep neural networks. In: Ceci, M., Hollmén, J., Todorovski, L., Vens, C., Džeroski, S. (eds.) ECML PKDD 2017. LNCS (LNAI), vol. 10534, pp. 187–202. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-71249-9_12
Wang, X., Xv, X., Wu, T.: The optimal reward baseline for policy-gradient reinforcement learning. Chin. J. Comput. Chin. Ed. 28(6), 1021–1026 (2005)
Cheng, S., Gu, R., Cheng, G.: Natural gradient reinforcement learning algorithm with TD(λ). Comput. Sci. 37(12), 186–189 (2010)
Luo, C., Su, R., Wang, X.: Adaptive stochastic parallel gradient descent algorithm and its application in coherent beam combining. Acta Optica Sinica 34(s1), s1010061–s1010065 (2014)
Zhang, L., Sun, H., Guo, H.: Auto focusing algorithm based on largest gray gradient summation. Acta Photonica Sinica 42(5), 605–610 (2013)
Wang, D., Zhang, J., Cao, W.: When will you arrive? Estimating travel time based on deep neural networks. In: Thirty-Second AAAI Conference on Artificial Intelligence, AAAI, vol. 32, pp. 2500-2507, New Orleans, Louisiana, USA (2018)
Pérez-GarcÃa, F., Sparks, R., Ourselin, S.: TorchIO: a python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning, 1–28. arXiv (2020)
Bi, B., Wang, Q., Coleman, J.: A novel mutation A212T in chloroplast Protoporphyrinogen oxidase (PPO1) confers resistance to PPO inhibitor Oxadiazon in Eleusine indica. Pest Manag. Sci. 76(5), 1786–1794 (2019)
Sharif, M., Attique, M., Tahir, M.Z.: A machine learning method with threshold based parallel feature fusion and feature selection for automated gait recognition. J. Organ. End User Comput. (JOEUC) 32(2), 67–92 (2020)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Lv, Z., Li, J., Xu, Z., Wang, Y., Li, H. (2021). Parallel Computing of Spatio-Temporal Model Based on Deep Reinforcement Learning. In: Liu, Z., Wu, F., Das, S.K. (eds) Wireless Algorithms, Systems, and Applications. WASA 2021. Lecture Notes in Computer Science(), vol 12937. Springer, Cham. https://doi.org/10.1007/978-3-030-85928-2_31
Download citation
DOI: https://doi.org/10.1007/978-3-030-85928-2_31
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-85927-5
Online ISBN: 978-3-030-85928-2
eBook Packages: Computer ScienceComputer Science (R0)