Sandeep P. Chinchali
ABSTRACT
We believe that most future video uploaded over the network will be consumed by machines for sensing tasks such as automated surveillance and mapping rather than for human consumption. Today's systems typically collect raw data from distributed sensors, such as drones, with the computer vision logic implemented in the cloud using deep neural networks (DNNs). They use standard video encoding techniques, send it over the network, and then decompress it at the cloud before using the vision DNN. In other words, data encoding and distribution is decoupled from the sensing goal. This is bandwidth inefficient because video encoding schemes, such as MPEG4, might send data tailored for human perception but irrelevant for the overall sensing goal.
We argue that data collection and distribution mechanisms should be co-designed with the eventual sensing objective. Specifically, we propose a distributed DNN architecture that learns end-to-end how to represent the raw sensor data and send it over the network such that it meets the eventual sensing task's needs. Such a design naturally adapts to varying network bandwidths between the sensors and the cloud, as well as automatically sends task-appropriate data features.
Supplemental Material
- R. Bellman. A markovian decision process. Technical report, DTIC Document, 1957.Google Scholar
- E. F. Camacho and C. B. Alba. Model predictive control. Springer Science & Business Media, 2013.Google Scholar
- C. Doersch. Tutorial on variational autoencoders. Available at: https://arxiv.org/abs/1606.05908, 2016.Google Scholar
- K. Fukushima. Neural network model for selective attention in visual pattern recognition and associative recall. Applied Optics, 26(23):4985--4992, 1987.Google ScholarCross Ref
- S. Han, H. Mao, and W. J. Dally. Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding. arXiv preprint arXiv:1510.00149, 2015.Google Scholar
- S. Han, J. Pool, J. Tran, and W. Dally. Learning both weights and connections for efficient neural network. In Advances in neural information processing systems, pages 1135--1143, 2015. Google ScholarDigital Library
- K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770--778, 2016.Google ScholarCross Ref
- W. Hess, D. Kohler, H. Rapp, and D. Andor. Real-time loop closure in 2d lidar slam. In Robotics and Automation (ICRA), 2016 IEEE International Conference on, pages 1271--1278. IEEE, 2016.Google ScholarDigital Library
- A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam. Mobilenets: Efficient convolutional neural networks for mobile vision applications. CoRR, abs/1704.04861, 2017.Google Scholar
- A. H. Jiang, D. L.-K. Wong, C. Canel, L. Tang, I. Misra, M. Kaminsky, M. A. Kozuch, P. Pillai, D. G. Andersen, and G. R. Ganger. Mainstream: Dynamic stem-sharing for multi-tenant video processing. In 2018 USENIX Annual Technical Conference (USENIX ATC 18), pages 29--42, Boston, MA, 2018. USENIX Association. Google ScholarDigital Library
- J. Jiang, G. Ananthanarayanan, P. Bodik, S. Sen, and I. Stoica. Chameleon: scalable adaptation of video analytics. In Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication, pages 253--266. ACM, 2018. Google ScholarDigital Library
- Y. Kang, J. Hauswald, C. Gao, A. Rovinski, T. Mudge, J. Mars, and L. Tang. Neurosurgeon: Collaborative intelligence between the cloud and mobile edge. ACM SIGPLAN Notices, 52(4):615--629, 2017. Google ScholarDigital Library
- J. H. Ko, T. Na, M. F. Amir, and S. Mukhopadhyay. Edge-host partitioning of deep neural networks with feature space encoding for resource-constrained internet-of-things platforms. CoRR, abs/1802.03835, 2018.Google Scholar
- Y. LeCun, Y. Bengio, and G. Hinton. Deep learning. nature, 521(7553):436, 2015.Google Scholar
- Y. LeCun, J. S. Denker, and S. A. Solla. Optimal brain damage. In Advances in neural information processing systems, pages 598--605, 1990. Google ScholarDigital Library
- Z. Liu, T. Liu, W. Wen, L. Jiang, J. Xu, Y. Wang, and G. Quan. Deepnjpeg: a deep neural network favorable jpeg-based image compression framework. In DAC, pages 18:1--18:6. ACM, 2018. Google ScholarDigital Library
- Z.-Q. Luo. Universal decentralized estimation in a bandwidth constrained sensor network. IEEE Transactions on information theory, 51(6):2210--2219, 2005. Google ScholarDigital Library
- H. Mao, R. Netravali, and M. Alizadeh. Neural adaptive video streaming with pensieve. In Proceedings of the Conference of the ACM Special Interest Group on Data Communication, pages 197--210. ACM, 2017. Google ScholarDigital Library
- V. Mnih, K. Kavukcuoglu, D. Silver, A. Graves, I. Antonoglou, D. Wierstra, and M. Riedmiller. Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602, 2013.Google Scholar
- J. Oueis and E. C. Strinati. Uplink traffic in future mobile networks: Pulling the alarm. In International Conference on Cognitive Radio Oriented Wireless Networks, pages 583--593. Springer, 2016.Google ScholarCross Ref
- A. Ribeiro and G. B. Giannakis. Bandwidth-constrained distributed estimation for wireless sensor networks-part i: Gaussian case. IEEE transactions on signal processing, 54(3):1131--1143, 2006. Google ScholarDigital Library
- D. Silver, G. Lever, N. Heess, T. Degris, D. Wierstra, and M. Riedmiller. Deterministic policy gradient algorithms. In T. Jebara and E. P. Xing, editors, Proceedings of the 31st International Conference on Machine Learning (ICML-14), pages 387--395. JMLR Workshop and Conference Proceedings, 2014. Google ScholarDigital Library
- N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research, 15(1):1929--1958, 2014. Google ScholarDigital Library
- R. Sutton and A. Barto. Reinforcement learning: an introduction. Neural Networks, IEEE Transactions on, 9(5):1054--1054, 1998. Google ScholarDigital Library
- C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich. Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1--9, 2015.Google ScholarCross Ref
- C. Szepesvári. Algorithms for reinforcement learning. Synthesis Lectures on Artificial Intelligence and Machine Learning, 4(1):1--103, 2010.Google ScholarCross Ref
- S. Teerapittayanon, B. McDanel, and H. T. Kung. Distributed deep neural networks over the cloud, the edge and end devices. In 37th IEEE International Conference on Distributed Computing Systems, ICDCS 2017, Atlanta, GA, USA, June 5-8, 2017, pages 328--339, 2017.Google ScholarCross Ref
- P. Vincent, H. Larochelle, I. Lajoie, Y. Bengio, and P.-A. Manzagol. Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion. J. Mach. Learn. Res., 11:3371--3408, Dec. 2010. Google ScholarDigital Library
- K. Xu, J. Ba, R. Kiros, K. Cho, A. Courville, R. Salakhudinov, R. Zemel, and Y. Bengio. Show, attend and tell: Neural image caption generation with visual attention. In International conference on machine learning, pages 2048--2057, 2015. Google ScholarDigital Library
- H. Yeo, S. Do, and D. Han. How will deep learning change internet video delivery? In Proceedings of the 16th ACM Workshop on Hot Topics in Networks, pages 57--64. ACM, 2017. Google ScholarDigital Library
Recommendations
Quantized neural networks: training neural networks with low precision weights and activations
We introduce a method to train Quantized Neural Networks (QNNs) -- neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At traintime the quantized weights and activations are used for computing the parameter ...
AQM performance in multiple congested link networks
ACM-SE 43: Proceedings of the 43rd annual Southeast regional conference - Volume 2TCP Reno, the most widely used implementation on the Internet, uses retransmission timeouts and the receipt of three duplicate ACKs to detect packet loss in the network. In case of multiple congested links, it is important to detect packet losses as ...
Granular neural networks
Fuzzy neural networks (FNNs) and rough neural networks (RNNs) both have been hot research topics in the artificial intelligence in recent years. The former imitates the human brain in dealing with problems, the other takes advantage of rough set theory ...
Comments