Minimax Policy for Heavy-Tailed Bandits

被引：3

作者：

Wei, Lai ^{[1
]}

Srivastava, Vaibhav ^{[1
]}

机构：

[1] Michigan State Univ, Dept Elect & Comp Engn, E Lansing, MI 48823 USA

来源：

IEEE CONTROL SYSTEMS LETTERS | 2021年 / 5卷 / 04期

关键词：

Heavy-tailed distribution; stochastic MAB; worst-case regret; minimax policy;

D O I：

10.1109/LCSYS.2020.3035767

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We study the stochastic Multi-Armed Bandit (MAB) problem under worst-case regret and heavytailed reward distribution. We modify the minimax policy MOSS for the sub-Gaussian reward distribution by using saturated empirical mean to design a new algorithm called Robust MOSS. We show that if the moment of order 1 + epsilon for the reward distribution exists, then the refined strategy has a worst-case regret matching the lower bound while maintaining a distribution-dependent logarithm regret.

引用

页码：1423 / 1428

页数：6

共 50 条

[21] Renewal reward processes with heavy-tailed inter-renewal times and heavy-tailed rewards
Levy, JB
Taqqu, MS
BERNOULLI, 2000, 6 (01) : 23 - 44
[22] Heavy-tailed log hydraulic conductivity distributions imply heavy-tailed log velocity distributions
Kohlbecker, MV
Wheatcraft, SW
Meerschaert, MM
WATER RESOURCES RESEARCH, 2006, 42 (04)
[23] On the Sample Complexity and Metastability of Heavy-tailed Policy Search in Continuous Control
Bedi, Amrit Singh
Parayil, Anjaly
Zhang, Junyu
Wang, Mengdi
Koppel, Alec
JOURNAL OF MACHINE LEARNING RESEARCH, 2024, 25
[24] Minimax Optimal Bandits for Heavy Tail Rewards
Lee, Kyungjae
Lim, Sungbin
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2022, 35 (04) : 4899 - 4901
[25] Heavy-tailed Independent Component Analysis
Anderson, Joseph
Goyal, Navin
Nandi, Anupama
Rademacher, Luis
2015 IEEE 56TH ANNUAL SYMPOSIUM ON FOUNDATIONS OF COMPUTER SCIENCE, 2015, : 290 - 309
[26] A PARAMETRIC BOOTSTRAP FOR HEAVY-TAILED DISTRIBUTIONS
Cornea-Madeira, Adriana
Davidson, Russell
ECONOMETRIC THEORY, 2015, 31 (03) : 449 - 470
[27] Estimating the Mean of Heavy-Tailed Distributions
Joachim Johansson
Extremes, 2003, 6 (2) : 91 - 109
[28] Search heuristics and heavy-tailed behaviour
Hulubei, T
O'Sullivan, B
PRINCIPLES AND PRACTICE OF CONSTRAINT PROGRAMMING - CP 2005, PROCEEDINGS, 2005, 3709 : 328 - 342
[29] Asymptotic Expansions for Heavy-Tailed Data
Pastor, Giancarlo
Mora-Jimenez, Inmaculada
Caamano, Antonio J.
Jantti, Riku
IEEE SIGNAL PROCESSING LETTERS, 2016, 23 (04) : 444 - 448
[30] HEAVY-TAILED BAYESIAN NONPARAMETRIC ADAPTATION
Agapiou, Sergios
Castillo, Ismael
ANNALS OF STATISTICS, 2024, 52 (04): : 1433 - 1459

← 1 2 3 4 5 →