Stein's Method Meets Computational Statistics: A Review of Some Recent Developments

被引:17
|
作者
Anastasiou, Andreas [1 ]
Barp, Alessandro [2 ]
Briol, Francois-Xavier [3 ]
Ebner, Bruno [4 ]
Gaunt, Robert E. [5 ]
Ghaderinezhad, Fatemeh [6 ]
Gorham, Jackson [7 ]
Gretton, Arthur [8 ]
Ley, Christophe [9 ]
Liu, Qiang [10 ]
Mackey, Lester [11 ]
Oates, Chris J. [12 ]
Reinert, Gesine [13 ]
Swan, Yvik [14 ]
机构
[1] Univ Cyprus, Dept Math & Stat, POB 20537, CY-1678 Nicosia, Cyprus
[2] Univ Cambridge, Engn Dept, Trumpington St, Cambridge CB2 1PZ, England
[3] UCL, 1 19 Torrington Pl, London WC1E 7HB, England
[4] Karlsruhe Inst Technol KIT, Inst Stochast, Englerstr 2, D-76128 Karlsruhe, Germany
[5] Univ Manchester, Oxford Rd, Manchester M13 9PL, England
[6] amfori, Gradient Bldg,Ave Tervueren 270, B-1150 Brussels, Belgium
[7] Whisper ai Inc, San Francisco, CA USA
[8] UCL, Sainsbury Wellcome Ctr, Gatsby Computat Neurosci Unit, 25 Howland St, London W1T 4JG, England
[9] Univ Luxembourg, Maison Nombre, 6 Ave Fonte, L-4364 Luxembourg, Luxembourg
[10] Univ Texas Austin, Austin, TX 78712 USA
[11] Microsoft Res New England, 1 Mem Dr, Cambridge, MA 02142 USA
[12] Newcastle Univ, Newcastle upon Tyne, England
[13] Univ Oxford, Dept Stat, 24-29 St Giles, Oxford OX1 3LB, England
[14] Univ Libre Bruxelles, Dept Math, CP 210,Blvd Triomphe, B-1050 Brussels, Belgium
基金
英国工程与自然科学研究理事会;
关键词
Stein's method; sample quality; approximate Markov chain Monte Carlo; variational inference; control variates; goodness-of-fit testing; maximum likelihood estimator; likelihood ratio; prior sensitiv-ity; MAXIMUM-LIKELIHOOD ESTIMATOR; VARIATIONAL GRADIENT DESCENT; OF-FIT TEST; NORMAL APPROXIMATION; MONTE-CARLO; WASSERSTEIN DISTANCE; GAMMA-DISTRIBUTION; GOODNESS; BOUNDS; CONVERGENCE;
D O I
10.1214/22-STS863
中图分类号
TU [建筑科学];
学科分类号
0813 ;
摘要
Stein's method compares probability distributions through the study of a class of linear operators called Stein operators. While mainly stud-ied in probability and used to underpin theoretical statistics, Stein's method has led to significant advances in computational statistics in recent years. The goal of this survey is to bring together some of these recent develop-ments, and in doing so, to stimulate further research into the successful field of Stein's method and statistics. The topics we discuss include tools to bench-mark and compare sampling methods such as approximate Markov chain Monte Carlo, deterministic alternatives to sampling methods, control variate techniques, parameter estimation and goodness-of-fit testing.
引用
收藏
页码:120 / 139
页数:20
相关论文
共 50 条