人工智能研究的新前线：生成式对抗网络  

The New Frontier of AI Research: Generative Adversarial Networks

作　　者：林懿伦 戴星原 李力[4] 王晓[1,3] 王飞跃 LIN Yi-Lun;DAI Xing-Yuan;LI Li;WANG Xiao;WANG Fei-Yue(The State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190;University of Chinese Academy of Sciences, Beijing 100049;Qingdao Academy of Intelligent Industries, Qingdao 266109;Department of Automation, Beijing National Research Center for Information Science and Technology （BNRist）, Tsinghua University, Beijing 100084;Research Center of Military Computational Experiments and Parallel System, National University of Defense Technology, Changsha 410073;Center of China Economic and Social Security, The University of Chinese Academy of Sciences, Beijing 101408)

机构地区：[1]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,北京100190 [2]中国科学院大学,北京100049 [3]青岛智能产业技术研究院,青岛266109 [4]北京信息科学与技术国家研究中心,清华大学自动化系,北京100084 [5]国防科学技术大学军事计算实验与平行系统技术中心,长沙410073 [6]中国科学院大学中国经济与社会安全研究中心,北京101408

出　　处：《自动化学报》2018年第5期775-792,共18页Acta Automatica Sinica

基　　金：国家自然科学基金(61533019,61702519); 北京市科技项目(D17110600030000,ZC179074Z)资助

摘　　要：生成式对抗网络(Generative adversarial networks,GAN)是当前人工智能学界最为重要的研究热点之一.其突出的生成能力不仅可用于生成各类图像和自然语言数据,还启发和推动了各类半监督学习和无监督学习任务的发展.本文概括了GAN的基本思想,并对近年来相关的理论与应用研究进行了梳理,总结了GAN常见的网络结构与训练方法,博弈形式,集成方法,并对一些应用场景进行了介绍.在此基础上,本文对GAN发展的内在逻辑进行了归纳总结.Recently, generative adversarial networks（GAN） have become one of the most popular topics in artificial intelligent field. Its outstanding capability of generating realistic samples not only revived the research of generative model, but also inspired the research of semi-supervised learning and unsupervised learning. In this paper, we introduce the basic idea of GAN, and comb its recent development in theory and practice. By concluding its improvements of network structures, optimization methods, the form of the game, the ensemble methods, and its applications, we found the inner logic of its development.

关 键 词：深度学习 生成式对抗网络 生成模型 对抗学习 平行学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]