近年資料量的成長和平行處理器的進步為深度學習(Deep learning)帶來了大幅度的成長，多組研究團隊運用深度學習在影像識別比賽中(ILSVRC)拿下第一名。然而，在訓練深度學習模型過程中需要大量的資料儲存空間和計算資源，對行動裝置而言有硬體上的限制，因此無法讓所有作業都在裝置上進行。
在這篇論文中我們探討如何運用行動雲端計算的架構，讓深度學習能夠在行動裝置上順利運作，並提出基本的架構和設計的系統流程。提出的系統除了深度學習的應用外，也負責處理模型的更新。我們在NVIDIA Jetson TK1開發版上運用設計的系統透過Faster R-CNN的架構實作了深度學習物件識別。系統的效能為1-4 FPS。

Deep learning has become a powerful technology in image recognition, gaming, information retrieval, and many other areas that need intelligent data processing. However, huge amount of data and complex computations prevent deep learning from being practical in mobile applications. In this thesis, we proposed a mobile cloud computing system for deep learning. The architecture puts the training process and model repository in cloud platforms, and the recognition process and data gathering in mobile devices. The communication is carried out via Git protocol to ensure the success of data transmission in unstable network environments. We used car camera object detection as an example application, and implemented the system on NVIDIA Jetson TK1. Experiment results show that detection rate can achieve 1 to 4 FPS with Faster R-CNN and ZF model, and the system can work well even when the network connection is unstable.

[1] http://www.image-net.org/challenges/LSVRC.
[2] http://www.mobilecloudcomputingforum.com.
[3] https://github.com/rbgirshick/py-faster-rcnn.
[4] Yoshua Bengio, Pascal Lamblin, Dan Popovici, Hugo Larochelle, et al. Greedy layer-wise training of deep networks. Advances in neural information processing systems, 19:153, 2007.
[5] Ronan Collobert and Jason Weston. A unified architecture for natural language processing: Deep neural networks with multitask learning. In Proceedings of the 25th international conference on Machine learning, pages 160–167. ACM, 2008.
[6] George E Dahl, Dong Yu, Li Deng, and Alex Acero. Context-dependent pretrained deep neural networks for large-vocabulary speech recognition. IEEE Transactions on Audio, Speech, and Language Processing, 20(1):30–42, 2012.
[7] Tim Dettmers. Deep learning in a nutshell, 2015. https://devblogs.nvidia.com/parallelforall/deep-learning-nutshell-history-training.
[8] Hoang T Dinh, Chonho Lee, Dusit Niyato, and Ping Wang. A survey of mobile cloud computing: architecture, applications, and approaches. Wireless communications and mobile computing, 13(18):1587–1611, 2013.
[9] Rasool Fakoor, Faisal Ladhak, Azade Nazi, and Manfred Huber. Using deep learning to enhance cancer diagnosis and classification. In Proceedings of the International Conference on Machine Learning, 2013.
[10] Ross Girshick. Fast r-cnn. In Proceedings of the IEEE International Conference on Computer Vision, pages 1440–1448, 2015.
[11] Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 580–587, 2014.
[12] Song Han, Huizi Mao, and William J Dally. Deep compression: Compressing deep neural network with pruning, trained quantization and huffman coding. CoRR, abs/1510.00149, 2, 2015.
[13] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In Proceedings of the IEEE International Conference on Computer Vision, pages 1026–1034, 2015.
[14] Geoffrey Hinton, Li Deng, Dong Yu, George E Dahl, Abdel-rahman Mohamed, Navdeep Jaitly, Andrew Senior, Vincent Vanhoucke, Patrick Nguyen, Tara N Sainath, et al. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal Processing Magazine, 29(6):82–97, 2012.
[15] Junxian Huang, Feng Qian, Alexandre Gerber, Z Morley Mao, Subhabrata Sen, and Oliver Spatscheck. A close examination of performance and power characteristics of 4g lte networks. In Proceedings of the 10th international conference on Mobile systems, applications, and services, pages 225–238. ACM, 2012.
[16] Yangqing Jia, Evan Shelhamer, Jeff Donahue, Sergey Karayev, Jonathan Long, Ross Girshick, Sergio Guadarrama, and Trevor Darrell. Caffe: Convolutional architecture for fast feature embedding. arXiv preprint arXiv:1408.5093, 2014.
[17] Justin Johnson, Andrej Karpathy, and Li Fei-Fei. Densecap: Fully convolutional localization networks for dense captioning. arXiv preprint arXiv:1511.07571, 2015.
[18] Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pages 1097–1105, 2012.
[19] Yann LeCun, LD Jackel, Leon Bottou, A Brunot, Corinna Cortes, JS Denker, Harris Drucker, I Guyon, UA Muller, Eduard Sackinger, et al. Comparison of learning algorithms for handwritten digit recognition. In International conference on artificial neural networks, volume 60, pages 53–60, 1995.
[20] Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. You only look once: Unified, real-time object detection. arXiv preprint arXiv:1506.02640, 2015.
[21] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster r-cnn: Towards real-time object detection with region proposal networks. In Advances in neural information processing systems, pages 91–99, 2015.
[22] Pierre Sermanet, David Eigen, Xiang Zhang, Micha¨el Mathieu, Rob Fergus, and Yann LeCun. Overfeat: Integrated recognition, localization and detection using convolutional networks. arXiv preprint arXiv:1312.6229, 2013.
[23] Richard Socher, Alex Perelygin, Jean Y Wu, Jason Chuang, Christopher D Manning, Andrew Y Ng, and Christopher Potts. Recursive deep models for semantic compositionality over a sentiment treebank. In Proceedings of the conference on empirical methods in natural language processing (EMNLP), volume 1631, page 1642. Citeseer, 2013.
[24] Nitish Srivastava, Geoffrey E Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 15(1):1929–1958, 2014.
[25] Ilya Sutskever, James Martens, George E Dahl, and Geoffrey E Hinton. On the importance of initialization and momentum in deep learning. ICML (3), 28:1139–1147, 2013.
[26] Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper with convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1–9, 2015.