Fig. 1 1 GAN Basic structure and computation procedure of GAN., z c c, InfoGAN, [7]., GAN,,. 2 Fig. 2 arxiv GAN Trend of the number of GAN pape

Size: px

Start display at page:

Download "Fig. 1 1 GAN Basic structure and computation procedure of GAN., z c c, InfoGAN, [7]., GAN,,. 2 Fig. 2 arxiv GAN Trend of the number of GAN pape"

袈倪
5 years ago
Views:

1 44 5 Vol. 44, No ACTA AUTOMATICA SINICA May, 2018 : 1, , 7.,,. (Generative adversarial networks, GAN),. 1 GAN GAN Goodfellow [1] GAN (Generator) (Discriminator), 1. p z ( ) z, G, G, G(z). G, z G(z) Manuscript received May 1, 2018 DOI /j.aas.2018.y ( , ) Supported by National Natural Science Foundation of China ( , ) :,,,,,. :., 2018, 44(5): Citation: Wang Kun-Feng, Zuo Wang-Meng, Tan Ying, Qin Tao, Li Li, Wang Fei-Yue. Generative adversarial networks: from generating data to creating intelligence. Acta Automatica Sinica, 2018, 44(5): The State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing Qingdao Academy of Intelligent Industries, Qingdao School of Computer Science and Technology, Harbin Institute of Technology, Harbin School of Electronics Engineering and Computer Science, Peking University, Beijing Microsoft Research Asia, Beijing Department of Automation, Tsinghua University, Beijing Research Center for Computational Experiments and Parallel Systems Technology, National University of Defense Technology, Changsha , G(z)., G(z) ( ), x. D G(z) x,,. G D,., G D,,. D G(z) x, G(z), x., G, G(z) p g x p data., D G, D G ;, G(z) x, G(z) x, D. Goodfellow [1] GAN,., GAN [2]. : (Mode collapse). G D, D, G(z) x, D, G, G. z, G p data, G(z) x D, G. GAN, [3]., Arjovsky [4] Wasserstein GAN, Earth-mover Jensen-Shannon,,. 2 arxiv GAN ( Generative adversarial networks Generative adversarial nets Adversarial

770 44 Fig. 1 1 GAN Basic structure and computation procedure of GAN., z c c, InfoGAN, [7].

2 arxiv GAN Trend of the number of GAN papers published on arxiv learning ), GAN.

lg stat.ml) (cs.cl) (cs.ai) (cs.sd) (eess.as) (cs.ro) (cs.cr) (cs.mm). 2 GAN: GAN G, ( ).

2 Fig. 1 1 GAN Basic structure and computation procedure of GAN., z c c, InfoGAN, [7]., GAN,,. 2 Fig. 2 arxiv GAN Trend of the number of GAN papers published on arxiv learning ), GAN., Goodfellow GAN,, 2016,.,.,, Zhu [5] CycleGAN,,., GAN. 3 arxiv GAN., (cs.cv), GAN ; (cs.lg stat.ml) (cs.cl) (cs.ai) (cs.sd) (eess.as) (cs.ro) (cs.cr) (cs.mm). 2 GAN: GAN G, ( ).,, GAN [6]., GAN,, 3 Fig. 3 arxiv GAN TOP 10 Top 10 subject categories of the GAN papers published on arxiv GAN, ( )., [8], [9], [10], [11], [12], [5], [13], [14]. ( ),., SimGAN,,, [12]., GAN,,., GAN,,, [15] ;,

3 5 : : 771 GAN, /, ; Mal- GAN,, [16]., GAN. 3 GAN GAN, [17]., ACP (Artificial systems, computational experiments, and parallel execution) [18].,.. [19 20] ACP.,,. GAN /, [12 13],,. [21]., ;,. GAN,,., GAN,. GAN 4 (GAN),,, GAN. 38., 13, GAN ; GAN ;., : GAN,, GAN,., GAN. GAN, (Energy reconstruction error GAN, E-REGAN)., ;,, GAN. MNIST CIFAR-10,, E-REGAN. GAN,,. GAN,,,.,,,.,,,,,.,.,.,., GAN,...

4 772 44,,., GAN,,.,, GAN,..,., (Low-rank generative adversarial network, LR-GAN). (Transform invariant low-rank textures, TILT),,.,, TILT,., LR-GAN.,,,., GAN,,,. GAN, GAN..,,, MNIST CIFAR-10,.,..,.,. Wasserstein GAN ;,...., GAN., GAN, GAN, GAN, (Ghost). (Maximum likelihood estimation, MLE),.,,,.,.,,,, MLE.,.,., GAN GAN,. GAN

5 5 : : 773,,. GAN,.,., Regression GAN (Regression GAN, RGAN), GAN.,.,.,,., RGAN.,.,. References 1 Goodfellow I J, Pouget-Abadie J, Mirza M, Xu B, Warde- Farley D, Ozair S, et al. Generative adversarial nets. In: Proceedings of the 27th International Conference on Neural Information Processing Systems. Montreal, Canada: Curran Associates, Inc., Creswell A, White T, Dumoulin V, Arulkumaran K, Sengupta B, Bharath A A. Generative adversarial networks: an overview. IEEE Signal Processing Magazine, 2018, 35(1): Wang Kun-Feng, Gou Chao, Duan Yan-Jie, Lin Yi-Lun, Zheng Xin-Hu, Wang Fei-Yue. Generative adversarial networks: the state of the art and beyond. Acta Automatica Sinica, 2017, 43(3): (,,,,,. GAN., 2017, 43(3): ) 4 Arjovsky M, Chintala S, Bottou L. Wasserstein GAN. arxiv preprint arxiv: , Zhu J Y, Park T, Isola P, Efros A A. Unpaired imageto-image translation using cycle-consistent adversarial networks. In: Proceedings of the 2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy: IEEE, Karras T, Aila T, Laine S, Lehtinen J. Progressive growing of GANs for improved quality, stability, and variation. arxiv preprint arxiv: , Chen X, Duan Y, Houthooft R, Schulman J, Sutskever I, Abbeel P. InfoGAN: interpretable representation learning by information maximizing generative adversarial nets. In: Proceedings of the 30th Conference on Neural Information Processing Systems. Barcelona, Spain: Curran Associates, Inc., Zhang H, Xu T, Li H S, Zhang S T, Huang X L, Wang X G, et al. StackGAN: text to photo-realistic image synthesis with stacked generative adversarial networks. arxiv preprint arxiv: , Zhu J Y, Krähenbühl P, Shechtman E, Efros A A. Generative visual manipulation on the natural image manifold. arxiv preprint arxiv: , Ledig C, Theis L, Huszar F, Caballero J, Cunningham A, Acosta A, et al. Photo-realistic single image superresolution using a generative adversarial network. arxiv preprint arxiv: , Santana E, Hotz G. Learning a driving simulator. arxiv preprint arxiv: , Shrivastava A, Pflster T, Tuzel O, Susskind J, Wang W D, Webb R. Learning from simulated and unsupervised images through adversarial training. In: Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA: IEEE, Liu M Y, Breuel T, Kautz J. Unsupervised image-to-image translation networks. In: Advances in Neural Information Processing Systems 30. Barcelona, Spain: Curran Associates, Inc., Wu J J, Zhang C K, Xue T F, Freeman B, Tenenbaum J. Learning a probabilistic latent space of object shapes via 3D generative-adversarial modeling. In: Advances in Neural Information Processing Systems 29. Barcelona, Spain: Curran Associates, Inc., Luc P, Couprie C, Chintala S, Verbeek J. Semantic segmentation using adversarial networks. arxiv preprint arxiv: , Hu W W, Tan Y. Generating adversarial malware examples for black-box attacks based on GAN. arxiv preprint arxiv: , Wang F Y, Wang X, Li L X, Li L. Steps toward parallel intelligence. IEEE/CAA Journal of Automatica Sinica, 2016, 3(4): Wang Fei-Yue. Parallel system methods for management and control of complex systems. Control and Decision, 2004, 19(5): , 514 (.., 2004, 19(5): , 514) 19 Wang Kun-Feng, Gou Chao, Wang Fei-Yue. Parallel vision: an ACP-based approach to intelligent vision computing. Acta Automatica Sinica, 2016, 42(10): (,,. : ACP., 2016, 42(10): ) 20 Wang K F, Gou C, Zheng N N, Rehg J M, Wang F Y. Parallel vision for perception and understanding of complex scenes: methods, framework, and perspectives. Artificial Intelligence Review, 2017, 48(3): Li Li, Lin Yi-Lun, Cao Dong-Pu, Zheng Nan-Ning, Wang Fei-Yue. Parallel learning a new framework for machine learning. Acta Automatica Sinica, 2017, 43(1): 1 8 (,,,,.., 2017, 43(1): 1 8)

774 自动化学报客座编委王坤峰中国科学院自动化研究所复杂系统管理与控制国家重点实验室副研究员. 主要研究方向为智能交通系统, 智能视觉计算, 机器学习. 本文通信作者. E-mail: kunfeng.wang@ia.ac.

His research interest covers intelligent transportation systems, intelligent visual computing, and machine learning. Corresponding author of this paper.

cn (ZUO Wang-Meng Professor at the School of Computer Science and Technology, Harbin Institute of Technology.

cn (TAN Ying Professor at the School of Electronics Engineering and Computer Science, Peking University.

E-mail: taoqin@microsoft.com (QIN Tao Senior researcher at Microsoft Research Asia.

E-mail: li-li@tsinghua.edu.cn (LI Li Associate professor at Department of Automation, Tsinghua University.

6 774 自动化学报客座编委王坤峰中国科学院自动化研究所复杂系统管理与控制国家重点实验室副研究员. 主要研究方向为智能交通系统, 智能视觉计算, 机器学习. 本文通信作者. (WANG Kun-Feng Associate professor at The State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences. His research interest covers intelligent transportation systems, intelligent visual computing, and machine learning. Corresponding author of this paper.) 左旺孟哈尔滨工业大学计算机科学与技术学院教授. 主要研究方向为计算机视觉, 机器学习, 生物特征识别. (ZUO Wang-Meng Professor at the School of Computer Science and Technology, Harbin Institute of Technology. His research interest covers computer vision, machine learning, and biometrics.) 谭营北京大学信息科学技术学院教授. 主要研究方向为计算智能, 群体智能, 机器学习. (TAN Ying Professor at the School of Electronics Engineering and Computer Science, Peking University. His research interest covers computational intelligence, swarm intelligence, and machine learning.) 44 卷秦涛微软亚洲研究院资深研究员, 主要研究方向为机器学习, 博弈论和多智能体系统, 搜索和在线广告. (QIN Tao Senior researcher at Microsoft Research Asia. His research interest covers machine learning, game theory and multi-agent system, search and advertising.) 李力清华大学自动化系副教授. 主要研究方向为人工智能和机器学习, 智能交通系统和智能汽车. (LI Li Associate professor at Department of Automation, Tsinghua University. His research interest covers artificial intelligence and machine learning, intelligent transportation systems and intelligent vehicles.) 王飞跃中国科学院自动化研究所复杂系统管理与控制国家重点实验室研究员. 国防科技大学军事计算实验与平行系统技术研究中心主任. 主要研究方向为智能系统和复杂系统的建模分析与控制. (WANG Fei-Yue Professor at The State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences. Director of the Research Center for Computational Experiments and Parallel Systems Technology, National University of Defense Technology. His research interest covers modeling, analysis, and control of intelligent systems and complex systems.)

,.,,,,., (Hyperledger), ( ), , The DAO,,.,,,,,,,, 2 : UDC AFC, ( ) ;, CyberSpace [2].,, (Cyberspace) (Cyber-physical-social systems, CPSS

,.,,,,., (Hyperledger), ( ), , The DAO,,.,,,,,,,, 2 : UDC AFC, ( ) ;, CyberSpace [2].,, (Cyberspace) (Cyber-physical-social systems, CPSS 43 9 Vol. 43, No. 9 2017 9 ACTA AUTOMATICA SINICA September, 2017 : 1, 2 3 4 5 1, 6, 7 1 :. 2015, ; 2016 7, Gartner, (Peak of inflated expectations); (Smart contract) (Smart asset) 2016,,, 2017-07-07 Manuscript