Conditional Generative Moment-Matching Networks

被引:0
作者
Ren, Yong
Li, Jialian
Luo, Yucen
Zhu, Jun [1 ]
机构
[1] Tsinghua Univ, Dept Comp Sci & Tech, TNList Lab, Beijing, Peoples R China
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 29 (NIPS 2016) | 2016年 / 29卷
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Maximum mean discrepancy (MMD) has been successfully applied to learn deep generative models for characterizing a joint distribution of variables via kernel mean embedding. In this paper, we present conditional generative moment-matching networks (CGMMN), which learn a conditional distribution given some input variables based on a conditional maximum mean discrepancy (CMMD) criterion. The learning is performed by stochastic gradient descent with the gradient calculated by back-propagation. We evaluate CGMMN on a wide range of tasks, including predictive modeling, contextual generation, and Bayesian dark knowledge, which distills knowledge from a Bayesian model by learning a relatively small CGMMN student network. Our results demonstrate competitive performance in all the tasks.
引用
收藏
页数:9
相关论文
共 50 条
[41]   A MOMENT-MATCHING METHOD TO STUDY THE VARIABILITY OF PHENOMENA DESCRIBED BY PARTIAL DIFFERENTIAL EQUATIONS [J].
Gerbeau, Jean-Frederic ;
Lombardi, Damiano ;
Tixier, Eliott .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2018, 40 (03) :B743-B765
[42]   AN IMPROVED MOMENT-MATCHING ALGORITHM FOR EVALUATING TOP-EVENT PROBABILITY BOUNDS [J].
TAKARAGI, K ;
SASAKI, R ;
SHINGAI, S .
IEEE TRANSACTIONS ON RELIABILITY, 1982, 31 (01) :45-48
[43]   A Linear Extension of Unscented Kalman Filter to Higher-Order Moment-Matching [J].
Liu, Jiang ;
Wang, Yujin ;
Zhang, Ju .
2014 IEEE 53RD ANNUAL CONFERENCE ON DECISION AND CONTROL (CDC), 2014, :5021-5026
[44]   Modified moment-matching method for estimating pointing parameters in the presence of atmospheric turbulence [J].
Zhang Yu ;
Li Xinyang ;
Rao Changhui .
APPLIED OPTICS, 2012, 51 (10) :C144-C151
[45]   Enhancing Stability in Training Conditional Generative Adversarial Networks via Selective Data Matching [J].
Kong, Kyeongbo ;
Kim, Kyunghun ;
Kang, Suk-Ju .
IEEE ACCESS, 2024, 12 :119647-119659
[46]   Efficient transient simulation of lossy packaging interconnects using moment-matching techniques [J].
Celik, M ;
Cangellaris, AC .
IEEE TRANSACTIONS ON COMPONENTS PACKAGING AND MANUFACTURING TECHNOLOGY PART B-ADVANCED PACKAGING, 1996, 19 (01) :64-73
[47]   Construction of Moment-matching Multinomial Lattices Using Vandermonde Matrices and Grobner Bases [J].
Lundengard, Karl ;
Ogutu, Carolyne ;
Silvestrov, Sergei ;
Ni, Ying ;
Weke, Patrick .
ICNPAA 2016 WORLD CONGRESS: 11TH INTERNATIONAL CONFERENCE ON MATHEMATICAL PROBLEMS IN ENGINEERING, AEROSPACE AND SCIENCES, 2017, 1798
[48]   Comment on "An algorithm for moment-matching scenario generation with application to financial portfolio optimization" [J].
Pablo Contreras, Juan ;
Bosch, Paul ;
Herrera, Mauricio .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2018, 269 (03) :1180-1184
[49]   On the use of moment-matching to build reduced order models in flexible multibody dynamics [J].
Michael Lehner ;
Peter Eberhard .
Multibody System Dynamics, 2006, 16 :191-211
[50]   Scenario Generation for Cooling, Heating, and Power Loads Using Generative Moment Matching Networks [J].
Liao, Wenlong ;
Wang, Yusen ;
Wang, Yuelong ;
Powell, Kody ;
Liu, Qi ;
Yang, Zhe .
CSEE JOURNAL OF POWER AND ENERGY SYSTEMS, 2022, 8 (06) :1730-1740