·182· 智能系统学报 第15卷 cient architecture search by network transformation[C]// [32]KOCSIS L,SZEPESVARI C.Bandit based monte-carlo Proceedings of the 32nd AAAI Conference on Artificial planning[C]//Proceedings of the 17th European Confer- Intelligence.New Orleans,USA,2018. ence on Machine Learning.Berlin,Germany,2006: [21]SUGANUMA M,SHIRAKAWA S,NAGAO T.A genet- 282-293. ic programming approach to designing convolutional [33]SANDHOLM T.Solving imperfect-information games[J] neural network architectures[C]//Proceedings of Genetic Science,2015,347(6218):122-123 and Evolutionary Computation Conference.Berlin,Ger- [34]RACANIERE S,WEBER T,REICHERT D P,et al.Ima- many,2017:497-504. gination-augmented agents for deep reinforcement learn- [22]CAI Han,YANG Jiacheng,ZHANG Weinan,et al.Path- ing[C]//Proceedings of the 31st Conference on Neural In- level network transformation for efficient architecture formation Processing Systems.Long Beach,USA,2017: search[Cl//Proceedings of the 35th International Confer- 5690-5701 ence on Machine Learning.Stockholmsmassan,Stock- [35]ZINKEVICH M.JOHANSON M.BOWLING M,et al. holm,Sweden.2018:677-686. Regret minimization in games with incomplete informa- [23]ELSKEN T,METZEN J H,HUTTER F.Efficient multi- tion[C]//Proceedings of the 20th International Conference objective neural architecture search via lamarckian evolu- on Neural Information Processing Systems.Red Hook, tion[Cl//Proceedings of 2019 International Conference on USA,2007:1729-1736. Learning Representations.New Orleans,USA,2019. [36]BROWN N.SANDHOLM T.Safe and nested subgame [24]ZOPH B,VASUDEVAN V,SHLENS J,et al.Learning solving for imperfect-information games[C]//Proceedings transferable architectures for scalable image recognition[C/ of the 31st Conference on Neural Information Processing Proceedings of 2018 IEEE/CVF Conference on Com- Systems.Long Beach,USA,2017:689-699. puter Vision and Pattern Recognition.Salt Lake City, [37]VINYALS O,BABUSCHKIN I,CZARNECKI W M,et USA.2018:8697-8710. al.Grandmaster level in StarCraft II using multi-agent re- [25]REAL E,AGGARWAL A,HUANG Yanping,et al.Reg- inforcement learning[J].Nature,2019,575(7782): ularized evolution for image classifier architecture 350-354. search[J].AAAI technical track:machine learning,2019, 作者简介： 33(1):4780-4789 蒋胤傑，博士研究生，主要研究方 [26]VON NEUMANN J,MORGENSTERN O,KUHN H W. 向为人工智能、神经网结构搜索。 et al.Theory of games and economic behavior[M].Prin- ceton:Princeton University Press.2007. [27]NASH JR J F.Equilibrium points in n-person games[]. Proceedings of the national academy of sciences of the United States of America,1950,36(1):48-49. [28]TANG Pingzhong.Reinforcement mechanism 况琨，助理教授，主要研究方向为 design[C]//Proceedings of the 26th International Joint 因果推理、稳定学习、可解释性机器学 Conference on Artificial Intelligence.Melbourne,Aus- 习以及AI在医学和法学的相关应 tralia.2017:5146-5150. 用。曾担任NIPS、AAAI、CIKM [29]PEROLAT J.LEIBO J Z,ZAMBALDI V,et al.A multi- ICDM等国际学术会议程序委员会委 agent reinforcement learning model of common-pool re- 员。发表10余篇顶级会议和期刊文 source appropriation[C]//Proceedings of the 31st Confer- 章，包括KDD、ICML、MM、AAAI ence on Neural Information Processing Systems.Long TKDD等。 Beach.USA.2017:3643-3652 [30]GOODFELLOW I J.POUGET-ABADIE J.MIRZA M. 吴飞，教授，博士生导师，浙江大 学人工智能研究所所长，担任中国图 et al.Generative adversarial nets[C]//Proceedings of the 象图形学学会第七届理事会理事、中 27th International Conference on Neural Information Pro- 国图象图形学学会动画与数字娱乐专 cessing Systems.Cambridge,USA,2014:2672-2680. 委会副主任、中国计算机学会多媒体 [31]SUTTON R S,MCALLESTER D,SINGH S,et al.Policy 技术专业委员会常务委员。主要研究 gradient methods for reinforcement learning with func- 方向为人工智能、跨媒体计算、多媒体 tion approximation[Cl//Proceedings of the 12th Interna- 分析与检索和统计学习理论。曾获宝钢优秀教师奖，“高校 tional Conference on Neural Information Processing Sys- 计算机专业优秀教师奖励计划”，教育部人工智能科技创新 ems.Cambridge,USA,1999:1057-1063. 专家组工作组组长。发表学术论文70余篇。cient architecture search by network transformation[C]// Proceedings of the 32nd AAAI Conference on Artificial Intelligence. New Orleans, USA, 2018. SUGANUMA M, SHIRAKAWA S, NAGAO T. A genet￾ic programming approach to designing convolutional neural network architectures[C]//Proceedings of Genetic and Evolutionary Computation Conference. Berlin, Ger￾many, 2017: 497–504. [21] CAI Han, YANG Jiacheng, ZHANG Weinan, et al. Path￾level network transformation for efficient architecture search[C]//Proceedings of the 35th International Confer￾ence on Machine Learning. Stockholmsmässan, Stock￾holm, Sweden, 2018: 677–686. [22] ELSKEN T, METZEN J H, HUTTER F. Efficient multi￾objective neural architecture search via lamarckian evolu￾tion[C]//Proceedings of 2019 International Conference on Learning Representations. New Orleans, USA, 2019. [23] ZOPH B, VASUDEVAN V, SHLENS J, et al. Learning transferable architectures for scalable image recognition[C]// Proceedings of 2018 IEEE/CVF Conference on Com￾puter Vision and Pattern Recognition. Salt Lake City, USA, 2018: 8697–8710. [24] REAL E, AGGARWAL A, HUANG Yanping, et al. Reg￾ularized evolution for image classifier architecture search[J]. AAAI technical track: machine learning, 2019, 33(1): 4780–4789. [25] VON NEUMANN J, MORGENSTERN O, KUHN H W, et al. Theory of games and economic behavior[M]. Prin￾ceton: Princeton University Press, 2007. [26] NASH JR J F. Equilibrium points in n-person games[J]. Proceedings of the national academy of sciences of the United States of America, 1950, 36(1): 48–49. [27] TANG Pingzhong. Reinforcement mechanism design[C]//Proceedings of the 26th International Joint Conference on Artificial Intelligence. Melbourne, Aus￾tralia, 2017: 5146–5150. [28] PÉROLAT J, LEIBO J Z, ZAMBALDI V, et al. A multi￾agent reinforcement learning model of common-pool re￾source appropriation[C]//Proceedings of the 31st Confer￾ence on Neural Information Processing Systems. Long Beach, USA, 2017: 3643–3652. [29] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Pro￾cessing Systems. Cambridge, USA, 2014: 2672–2680. [30] SUTTON R S, MCALLESTER D, SINGH S, et al. Policy gradient methods for reinforcement learning with func￾tion approximation[C]//Proceedings of the 12th Interna￾tional Conference on Neural Information Processing Sys￾tems. Cambridge, USA, 1999: 1057–1063. [31] KOCSIS L, SZEPESVÁRI C. Bandit based monte-carlo planning[C]//Proceedings of the 17th European Confer￾ence on Machine Learning. Berlin, Germany, 2006: 282–293. [32] SANDHOLM T. Solving imperfect-information games[J]. Science, 2015, 347(6218): 122–123. [33] RACANIÈRE S, WEBER T, REICHERT D P, et al. Ima￾gination-augmented agents for deep reinforcement learn￾ing[C]//Proceedings of the 31st Conference on Neural In￾formation Processing Systems. Long Beach, USA, 2017: 5690–5701. [34] ZINKEVICH M, JOHANSON M, BOWLING M, et al. Regret minimization in games with incomplete informa￾tion[C]//Proceedings of the 20th International Conference on Neural Information Processing Systems. Red Hook, USA, 2007: 1729–1736. [35] BROWN N, SANDHOLM T. Safe and nested subgame solving for imperfect-information games[C]//Proceedings of the 31st Conference on Neural Information Processing Systems. Long Beach, USA, 2017: 689–699. [36] VINYALS O, BABUSCHKIN I, CZARNECKI W M, et al. Grandmaster level in StarCraft II using multi-agent re￾inforcement learning[J]. Nature, 2019, 575(7782): 350–354. [37] 作者简介： 蒋胤傑，博士研究生，主要研究方 向为人工智能、神经网络结构搜索。 况琨，助理教授，主要研究方向为 因果推理、稳定学习、可解释性机器学 习以及 AI 在医学和法学的相关应 用。曾担任 NIPS、AAAI、CIKM、 ICDM 等国际学术会议程序委员会委 员。发表 10 余篇顶级会议和期刊文 章，包括 KDD、ICML、MM、AAAI、 TKDD 等。 吴飞，教授，博士生导师，浙江大 学人工智能研究所所长，担任中国图 象图形学学会第七届理事会理事、中 国图象图形学学会动画与数字娱乐专 委会副主任、中国计算机学会多媒体 技术专业委员会常务委员。主要研究 方向为人工智能、跨媒体计算、多媒体 分析与检索和统计学习理论。曾获宝钢优秀教师奖，“高校 计算机专业优秀教师奖励计划”，教育部人工智能科技创新 专家组工作组组长。发表学术论文 70 余篇。 ·182· 智 能 系 统 学 报 第 15 卷