Variational approximation for importance sampling

被引：0
|
作者：
Xiao Su
Yuguo Chen
机构：
[1] Amazon,Department of Statistics
[2] University of Illinois at Urbana-Champaign,undefined
来源：
Computational Statistics | 2021年 / 36卷
关键词：
-divergence; Monte Carlo; Proposal distribution; Variational inference;
D O I：
暂无
中图分类号：
学科分类号：
摘要：
We propose an importance sampling algorithm with proposal distribution obtained from variational approximation. This method combines the strength of both importance sampling and variational method. On one hand, this method avoids the bias from variational method. On the other hand, variational approximation provides a way to design the proposal distribution for the importance sampling algorithm. Theoretical justification of the proposed method is provided. Numerical results show that using variational approximation as the proposal can improve the performance of importance sampling and sequential importance sampling.
引用
收藏
页码：1901 / 1930
页数：29
相关论文
共 50 条
  • [41] Active Importance Sampling for Variational Objectives Dominated by Rare Events: Consequences for Optimization and Generalization
    Rotskoff, Grant M.
    Mitchell, Andrew R.
    Vanden-Eijnden, Eric
    MATHEMATICAL AND SCIENTIFIC MACHINE LEARNING, VOL 145, 2021, 145 : 757 - 780
  • [42] Variational Monte Carlo calculation of two body properties in atoms:: Importance sampling considerations
    Sarsa, A
    Gálvez, FJ
    Buendía, E
    COMPUTER PHYSICS COMMUNICATIONS, 1999, 121 : 493 - 495
  • [43] Variational Monte Carlo calculation of two body properties in atoms: Importance sampling considerations
    Sarsa, Antonio
    Gálvez, Francisco J.
    Buendía, Enrique
    Computer Physics Communications, 1999, 121 : 493 - 495
  • [44] Variational shape approximation
    Cohen-Steiner, D
    Alliez, P
    Desbrun, M
    ACM TRANSACTIONS ON GRAPHICS, 2004, 23 (03): : 905 - 914
  • [45] Variational Rejection Sampling
    Grover, Aditya
    Gummadi, Ramki
    Lazaro-Gredilla, Miguel
    Schuurmans, Dale
    Ermon, Stefano
    INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 84, 2018, 84
  • [46] Adaptive importance sampling for value function approximation in off-policy reinforcement learning
    Hachiya, Hirotaka
    Akiyama, Takayuki
    Sugiayma, Masashi
    Peters, Jan
    NEURAL NETWORKS, 2009, 22 (10) : 1399 - 1410
  • [47] Comparison of a simple approximation for multinormal integration with an importance sampling-based simulation method
    Pandey, MD
    Sarkar, A
    PROBABILISTIC ENGINEERING MECHANICS, 2002, 17 (02) : 215 - 218
  • [48] Moderate Deviations for the Optimal Values of Sample Average Approximation with Adaptive Multiple Importance Sampling
    Wenjin ZHANG
    Journal of Mathematical Research with Applications, 2025, 45 (02) : 275 - 284
  • [49] An adaptive zero-variance importance sampling approximation for static network dependability evaluation
    Tuffin, Bruno
    Saggadi, Samira
    L'Ecuyer, Pierre
    COMPUTERS & OPERATIONS RESEARCH, 2014, 45 : 51 - 59
  • [50] An Adaptive Zero-Variance Importance Sampling Approximation for Static Network Dependability Evaluation
    L'Ecuyer, Pierre
    Saggadi, Samira
    Tuffin, Bruno
    SNA + MC 2013 - JOINT INTERNATIONAL CONFERENCE ON SUPERCOMPUTING IN NUCLEAR APPLICATIONS + MONTE CARLO, 2014,
12345