臺灣博碩士論文加值系統

針對行動及嵌入式應用的壓縮卷積神經網路架構在先前的研究中已被廣泛提出。然而受限於卷積神經網路靜態的特性，大部分的方法仍需要使用者訓練數個複雜度相異的模型來應付動態的運算限制環境。本論文提出了一個能夠動態調節給定模型計算量的框架。這個框架會在指定的運算資源限制下，選擇預訓練模型中合適的部件(區塊或是途徑) 來進行輸入圖片的鑑定。更具體一點來說，我們提出了一個利用演員─評論家演算法的強化學習方法，訓練一個代理模型在不更動具有支狀結構的預訓練模型架構下，限制模型進行推論時內部元件的使用量。我們使用殘差網路架構作為預訓練模型，並透過CIFAR10以及CIFAR100資料集上的測試來檢測我們所提出的框架。實驗結果顯示我們的框架可以在僅使用單一模型的情況下，於大範圍的計算量預算限制中有效的將模型推論所需要的運算量與推論後得到的分類準確度進行平衡。透過我們提出的方法，在基於ResNeXt29(4×16d) 的預訓練模型上能將其運算量根據使用者的需求降至原先的25%，同時也能夠在此運算變動範圍中於CIFAR10資料集上得到82.5% 至93.4%的分類準確率。

Previous works have shown several approaches to obtain compact convolutional neural networks for mobile and embedded applications. However, due to the static nature of most convolutional neural networks, these approaches require training an individual model for each different budget constraints. In this thesis, we introduce BudgetNet, a framework that dynamically regulates the computational cost of a given model during the inference phase. Our framework directly selects which components, blocks or paths, of a pretrained model to evaluate a given image under assigned budget constraint. Especially, we propose a reinforcement learning approach with actor-critic algorithms to train an agent that dynamically limits the component usage for each pretrained model consisted of branch structures without modifying their architecture. We validate our proposed framework with ResNet and ResNeXt on CIFAR10 and CIFAR100 respectively. The results show that, with a single model, our method efficiently trades off between computation cost and accuracy over a wide range of budget constraints. Based on the ResNeXt29 (4×16d) model, our method can control the amount of computation from lowest 25 % towards original computation according to user’s request while achieving 82.5% to 93.4% accuracy on CIFAR10.

誌謝iii
摘要v
Abstract vii
1 Introduction 1
2 Related Work 5
2.1 Model Compression 5
2.2 Efficient Network Architecture 6
2.3 Dynamic Network 6
3 Approach 9
3.1 Pre-trained Models with Branch Structure 9
3.2 Actor-Critic Method for Dynamic Inference 10
3.2.1 Policy Function and Value Function 11
3.2.2 Reward Function 12
3.2.3 Optimization and Back-propagation 13
3.3 Training BudgetNet 14
3.3.1 Encouraging Exploration 15
3.3.2 Curriculum Learning and Joint-Finetuning 15
4 Experiment 19
4.1 Experimental Setup 19
4.1.1 Datasets 19
4.1.2 Pretraind Models 19
4.1.3 Actor-Critic Architecture 22
4.2 Experiment Results 23
5 Conclusion Remarks 29
Bibliography 31

[1] P. Auer, N. CesaBianchi, and P. Fischer. Finitetime analysis of the multiarmed bandit problem. Machine learning, 47:235–256, 2002.
[2] Y. Bengio, J. Louradour, R. Collobert, and J. Weston. Curriculum learning. In Proceedings of the 26th annual international conference on machine learning, pages41–48. ACM, 2009.
[3] M. Bojarski, D. Del Testa, D. Dworakowski, B. Firner, B. Flepp, P. Goyal, L. D. Jackel, M. Monfort, U. Muller, J. Zhang, et al. End to end learning for self-driving cars. arXiv preprint arXiv:1604.07316, 2016.
[4] P. W. Glynn. Likelihood ratio gradient estimation for stochastic systems. Communications of the ACM, 33(10):75–84, 1990.
[5] D. Gmach, J. Rolia, L. Cherkasova, and A. Kemper. Workload analysis and demand prediction of enterprise data center applications. In 2007 IEEE 10th International Symposium on Workload Characterization, pages 171–180. IEEE, 2007.
[6] I. Grondman, L. Busoniu, G. A. Lopes, and R. Babuska. A survey of actorcritic reinforcement learning: Standard and natural policy gradients. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 42(6):1291–1307, 2012.
[7] S. Gupta, A. Agrawal, K. Gopalakrishnan, and P. Narayanan. Deep learning with limited numerical precision. In International Conference on Machine Learning, pages 1737–1746, 2015.
[8] 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.
[9] 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.
[10] G. Hinton, O. Vinyals, and J. Dean. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531, 2015.
[11] 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. arXiv preprint arXiv:1704.04861, 2017.
[12] G. Hu, Y. Yang, D. Yi, J. Kittler, W. Christmas, S. Z. Li, and T. Hospedales. When face recognition meets with deep learning: an evaluation of convolutional neural networks for face recognition. In Proceedings of the IEEE international conference on computer vision workshops, pages 142–150, 2015.
[13] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger. Densely connected convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 4700–4708, 2017.
[14] F. N. Iandola, S. Han, M. W. Moskewicz, K. Ashraf, W. J. Dally, and K. Keutzer. Squeezenet: Alexnetlevel accuracy with 50x fewer parameters and< 0.5 mb model size. arXiv preprint arXiv:1602.07360, 2016.
[15] B. Jacob, S. Kligys, B. Chen, M. Zhu, M. Tang, A. Howard, H. Adam, and D. Kalenichenko. Quantization and training of neural networks for efficient integerarithmeticonly inference. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 2704–2713, 2018.
[16] M. Jaderberg, A. Vedaldi, and A. Zisserman. Speeding up convolutional neural networks with low rank expansions. arXiv preprint arXiv:1405.3866, 2014.
[17] V. R. Konda and J. N. Tsitsiklis. Actorcritic algorithms. In Advances in neural information processing systems, pages 1008–1014, 2000.
[18] A. Krizhevsky et al. Learning multiple layers of features from tiny images. Technical report, Citeseer, 2009.
[19] A. Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classification with deep convolutional neural networks. In Proceedings of the 25th International Conference on Neural Information Processing Systems Volume 1, NIPS’12, pages 1097–1105, USA, 2012.
[20] S. Lee, S. Chang, and N. Kwak. Urnet : Userresizable residual networks with conditional gating module. arXiv preprint arXiv:1901.04687, 2019.
[21] I. Lenz, H. Lee, and A. Saxena. Deep learning for detecting robotic grasps. The International Journal of Robotics Research, 34(45): 705–724, 2015.
[22] H. Li, A. Kadav, I. Durdanovic, H. Samet, and H. P. Graf. Pruning filters for efficient convnets. arXiv preprint arXiv:1608.08710, 2016.
[23] T. Li, J. Zhang, S. Y. Philip, Y. Zhang, and Y. Yan. Deep dynamic network embedding for link prediction. IEEE Access, 6:29219–29230, 2018.
[24] L. Liu and J. Deng. Dynamic deep neural networks: Optimizing accuracyefficiency tradeoffs by selective execution. In ThirtySecond AAAI Conference on Artificial Intelligence, 2018.
[25] B. McDanel, S. Teerapittayanon, and H. Kung. Incomplete dot products for dynamic computation scaling in neural network inference. In 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA), pages 186–193. IEEE, 2017.
[26] M. Mirza and S. Osindero. Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784, 2014.
[27] V. Mnih, A. P. Badia, M. Mirza, A. Graves, T. Lillicrap, T. Harley, D. Silver, and K. Kavukcuoglu. Asynchronous methods for deep reinforcement learning. In International conference on machine learning, pages 1928–1937, 2016.
[28] 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.
[29] A. Mtibaa, A. Fahim, K. A. Harras, and M. H. Ammar. Towards resource sharing in mobile device clouds: Power balancing across mobile devices. In ACM SIGCOMM Computer Communication Review, volume 43, pages 51–56. ACM, 2013.
[30] A. Romero, N. Ballas, S. E. Kahou, A. Chassang, C. Gatta, and Y. Bengio. Fitnets: Hints for thin deep nets. arXiv preprint arXiv:1412.6550, 2014.
[31] K. Simonyan and A. Zisserman. Very deep convolutional networks for largescale image recognition. In International Conference on Learning Representations, 2015.
[32] S. Srinivas and R. V. Babu. Datafree parameter pruning for deep neural networks. arXiv preprint arXiv:1507.06149, 2015.
[33] R. S. Sutton and A. G. Barto. Reinforcement Learning: An Introduction. MIT Press, 1998.
[34] R. S. Sutton, D. A. McAllester, S. P. Singh, and Y. Mansour. Policy gradient methods for reinforcement learning with function approximation. In Advances in neural information processing systems, pages 1057–1063, 2000.
[35] C. Tai, T. Xiao, Y. Zhang, X. Wang, et al. Convolutional neural networks with lowrank regularization. arXiv preprint arXiv:1511.06067, 2015.
[36] R. Teja Mullapudi, W. R. Mark, N. Shazeer, and K. Fatahalian. Hydranets: Specialized dynamic architectures for efficient inference. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8080–8089, 2018.
[37] A. Veit, M. J. Wilber, and S. Belongie. Residual networks behave like ensembles of relatively shallow networks. In Advances in neural information processing systems, pages 550–558, 2016.
[38] J. Wu, C. Leng, Y. Wang, Q. Hu, and J. Cheng. Quantized convolutional neural networks for mobile devices. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4820–4828, 2016.
[39] Z. Wu, T. Nagarajan, A. Kumar, S. Rennie, L. S. Davis, K. Grauman, and R. Feris. Blockdrop: Dynamic inference paths in residual networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8817–8826, 2018.
[40] S. Xie, R. Girshick, P. Dollár, Z. Tu, and K. He. Aggregated residual transformations for deep neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1492–1500, 2017.
[41] X. Yu, T. Liu, X. Wang, and D. Tao. On compressing deep models by low rank and sparse decomposition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 7370–7379, 2017.
[42] X. Zhang, X. Zhou, M. Lin, and J. Sun. Shufflenet: An extremely efficient convolutional neural network for mobile devices. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6848–6856, 2018.
[43] B. Zoph, V. Vasudevan, J. Shlens, and Q. V. Le. Learning transferable architectures for scalable image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8697–8710, 2018.