Convergence Guarantees for Adaptive Bayesian Quadrature Methods

被引:0
|
作者
Kanagawa, Motonobu [1 ,2 ,3 ]
Hennig, Philipp [2 ,3 ]
机构
[1] EURECOM, Sophia Antipolis, France
[2] Univ Tubingen, Tubingen, Germany
[3] Max Planck Inst Intelligent Syst, Tubingen, Germany
基金
欧洲研究理事会;
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
Adaptive Bayesian quadrature (ABQ) is a powerful approach to numerical integration that empirically compares favorably with Monte Carlo integration on problems of medium dimensionality (where non-adaptive quadrature is not competitive). Its key ingredient is an acquisition function that changes as a function of previously collected values of the integrand. While this adaptivity appears to be empirically powerful, it complicates analysis. Consequently, there are no theoretical guarantees so far for this class of methods. In this work, for a broad class of adaptive Bayesian quadrature methods, we prove consistency, deriving non-tight but informative convergence rates. To do so we introduce a new concept we call weak adaptivity. Our results identify a large and flexible class of adaptive Bayesian quadrature rules as consistent, within which practitioners can develop empirically efficient methods.
引用
收藏
页数:12
相关论文
共 50 条
  • [41] Layer-Wise Adaptive Gradient Sparsification for Distributed Deep Learning with Convergence Guarantees
    Shi, Shaohuai
    Tang, Zhenheng
    Wang, Qiang
    Zhao, Kaiyong
    Chu, Xiaowen
    ECAI 2020: 24TH EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE, 2020, 325 : 1467 - 1474
  • [42] IMPROVEMENTS OF ADAPTIVE NEWTON-COTES QUADRATURE METHODS.
    Ninomiya, Ichizo
    Journal of Information Processing, 1980, 3 (03) : 162 - 170
  • [43] Layer adaptive node selection in Bayesian neural networks: Statistical guarantees and implementation details
    Jantre, Sanket
    Bhattacharya, Shrijita
    Maiti, Tapabrata
    NEURAL NETWORKS, 2023, 167 : 309 - 330
  • [44] Convergence of adaptive biasing potential methods for diffusions
    Benaim, Michel
    Brehier, Charles-Edouard
    COMPTES RENDUS MATHEMATIQUE, 2016, 354 (08) : 842 - 846
  • [45] Adaptive Finite Element Methods with convergence rates
    Peter Binev
    Wolfgang Dahmen
    Ron DeVore
    Numerische Mathematik, 2004, 97 : 219 - 268
  • [46] ON THE CONVERGENCE OF ADAPTIVE SEQUENTIAL MONTE CARLO METHODS
    Beskos, Alexandros
    Jasra, Ajay
    Kantas, Nikolas
    Thiery, Alexandre
    ANNALS OF APPLIED PROBABILITY, 2016, 26 (02): : 1111 - 1146
  • [47] On the convergence of adaptive iterative linearized Galerkin methods
    Heid, Pascal
    Wihler, Thomas P.
    CALCOLO, 2020, 57 (03)
  • [48] CORRIGENDUM TO "CONVERGENCE OF ADAPTIVE, DISCONTINUOUS GALERKIN METHODS"
    Kreuzer, Christian
    Georgoulis, Emmanuil H.
    MATHEMATICS OF COMPUTATION, 2021, 90 (328) : 637 - 640
  • [49] Adaptive finite element methods with convergence rates
    Binev, P
    Dahmen, W
    DeVore, R
    NUMERISCHE MATHEMATIK, 2004, 97 (02) : 219 - 268
  • [50] Adaptive boundary element methods with convergence rates
    Tsogtgerel Gantumur
    Numerische Mathematik, 2013, 124 : 471 - 516