Evaluating Edge-Cloud Computing Trade-Offs for Mobile Object Detection and Classification with Deep Learning
Keywords:deep neural networks, edge-cloud collaboration, image classification, object detection
Internet-of-Things (IoT) applications based on Artificial Intelligence, such as mobile object detection and recognition from images and videos, may greatly benefit from inferences made by state-of-the-art Deep Neural Network(DNN) models. However, adopting such models in IoT applications poses an important challenge since DNNs usually require lots of computational resources (i.e. memory, disk, CPU/GPU, and power), which may prevent them to run on resource-limited edge devices. On the other hand, moving the heavy computation to the Cloud may significantly increase running costs and latency of IoT applications. Among the possible strategies to tackle this challenge are: (i) DNN model partitioning between edge and cloud; and (ii) running simpler models in the edge and more complex ones in the cloud, with information exchange between models, when needed. Variations of strategy (i) also include: running the entire DNN on the edge device (sometimes not feasible) and running the entire DNN on the cloud. All these strategies involve trade-offs in terms of latency, communication, and financial costs. In this article we investigate such trade-offs in real-world scenarios. We conduct several experiments using deep learning models for image-based object detection and classification. Our setup includes a Raspberry PI 3 B+ and a cloud server equipped with a GPU. Different network bandwidths are also evaluated. Our results provide useful insights about the aforementioned trade-offs. The partitioning experiment showed that, overall, running the inferences entirely on the edge or entirely on the cloud server are the best options. The collaborative approach yielded a significant increase in accuracy without penalizing running costs too much.
Abadi, M. et al.TensorFlow: Large-scale machine learning on heterogeneous systems, 2015. Software available fromtensorflow.org.
Bengio, Y., Courville, A., and Vincent, P. Representation learning: A review and new perspectives. IEEE transactions on pattern analysis and machine intelligence 35 (8): 1798–1828, 2013.
Carneiro, S. A., da Silva, G. P., Guimaraes, S. J. F.,and Pedrini, H.Fight detection in video sequences basedon multi-stream convolutional neural networks. In 2019 32nd SIBGRAPI Conference on Graphics,Patterns and Images (SIBGRAPI). IEEE, pp. 8–15, 2019.
Chollet, F. Xception: Deep learning with depthwise separable convolutions, 2016.
Chollet, F. et al. Keras. https://keras.io, 2015.
da Silva Marques, R. C., Serra, A. C., França, J. V. F., Diniz, J. O. B., Junior, G. B., de Almeida, J. D. S., da Silva, M. I. A., and Monteiro, E. M. G. Image-based electric consumption recognition via multi-tasklearning. In 2019 8th Brazilian Conference on Intelligent Systems (BRACIS). IEEE, pp. 419–424, 2019.
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. ImageNet: A Large-Scale Hierarchical Image Database. InCVPR09, 2009.
dos Santos, F. P. and Ponti, M. A. Robust feature spaces from pre-trained deep network layers for skin lesion classification. In 2018 31st SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI). IEEE, pp. 189–196, 2018.
Eshratifar, A. E., Esmaili, A., and Pedram, M. Towards collaborative intelligence friendly architectures for deeplearning. In 20th International Symposium on Quality Electronic Design (ISQED). IEEE, pp. 14–19, 2019.
Gill, Z. User-driven collaborative intelligence: social networks as crowdsourcing ecosystems. In CHI’12 Extended Abstracts on Human Factors in Computing Systems. pp. 161–170, 2012.
Hadidi, R., Cao, J., Ryoo, M. S., and Kim, H. Collaborative execution of deep neural networks on internet of things devices. CoRR vol. abs/1901.02537, 2019.
Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., and Adam,H. Mobilenets: Efficient convolutional neural networks for mobile vision applications, 2017.
Kang, Y., Hauswald, J., Gao, C., Rovinski, A., Mudge, T., Mars, J., and Tang, L. Neurosurgeon: Collaborative intelligence between the cloud and mobile edge. In Proceedings of the Twenty-Second International Conference on Architectural Support for Programming Languages and Operating Systems. ACM, Xi’an, China, pp. 615–629, 2017.
Khan, A., Sohail, A., Zahoora, U., and Qureshi, A. S. A survey of the recent architectures of deep convolutional neural networks. Artificial Intelligence Review, 2020.
LeCun, Y., Bengio, Y., and Hinton, G. Deep learning. nature 521 (7553): 436–444, 2015.
Liu, S., Liu, L., Tang, J., Yu, B., Wang, Y., and Shi, W. Edge computing for autonomous driving: Opportunities and challenges. Proceedings of the IEEE 107 (8): 1697–1716, 2019.
Magalhães, W., Gomes, H., Marinho, L., Aguiar, G., and Silveira, P. Investigating mobile edge-cloud trade-offsof object detection with yolo. In Anais do VII Symposium on Knowledge Discovery, Mining and Learning. SBC, pp.49–56, 2019.
Marinus, J. V. d. M. L., de Araújo, J. M. F. R., and Gomes, H. M. Reconstructed phase space and convolutional neural networks for classifying voice pathologies. In Iberoamerican Congress on Pattern Recognition. Springer, pp.792–801, 2018.
Misra, A., Jayarajah, K., Weerakoon, D., Tandriansyah, R., Yao, S., and Abdelzaher, T. Dependable machine intelligence at the tactical edge. In Artificial Intelligence and Machine Learning for Multi-Domain Operations Applications. Vol. 11006. International Society for Optics and Photonics, pp. 1100608, 2019.
Pasandi, H. B. and Nadeem, T. Convince: Collaborative cross-camera video analytics at the edge. arXiv preprint arXiv:2002.03797, 2020.
Ponti, M. A., Ribeiro, L. S. F., Nazare, T. S., Bui, T., and Collomosse, J. Everything you wanted to know about deep learning for computer vision but were afraid to ask. In 2017 30th SIBGRAPI conference on graphics, patterns and images tutorials (SIBGRAPI-T). IEEE, pp. 17–41, 2017.
Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 779–788, 2016.
Redmon, J. and Farhadi, A. YOLO9000: better, faster, stronger. CoRR vol. abs/1612.08242, 2016.
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., and Chen, L.-C. Mobilenetv2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 4510–4520, 2018.
Shi, W., Hou, Y., Zhou, S., Niu, Z., Zhang, Y., and Geng, L. Improving device-edge cooperative inference of deep learning via 2-step pruning. CoRR vol. abs/1903.03472, 2019.
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A. Going deeper with convolutions, 2014.
Teerapittayanon, S., McDanel, B., and Kung, H. T. Distributed deep neural networks over the cloud, the edge and end devices. In 2017 IEEE 37th International Conference on Distributed Computing Systems(ICDCS). Institute of Electrical and Electronics Engineers (IEEE), Atlanta, USA, pp. 328–339, 2017.
Weerakoon, D., Jayarajah, K., and Misra, A. Resilient collaborative intelligence for adversarial iot environments, 2019.