Saved in:
Bibliographic Details
Main Authors: Liu, Yueyang, Kuang, Xu, Van Roy, Benjamin
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2205.01970
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908347371880448
author Liu, Yueyang
Kuang, Xu
Van Roy, Benjamin
author_facet Liu, Yueyang
Kuang, Xu
Van Roy, Benjamin
contents Thompson sampling has proven effective across a wide range of stationary bandit environments. However, as we demonstrate in this paper, it can perform poorly when applied to non-stationary environments. We attribute such failures to the fact that, when exploring, the algorithm does not differentiate actions based on how quickly the information acquired loses its usefulness due to non-stationarity. Building upon this insight, we propose predictive sampling, an algorithm that deprioritizes acquiring information that quickly loses usefulness. A theoretical guarantee on the performance of predictive sampling is established through a Bayesian regret bound. We provide versions of predictive sampling for which computations tractably scale to complex bandit environments of practical interest. Through numerical simulations, we demonstrate that predictive sampling outperforms Thompson sampling in all non-stationary environments examined.
format Preprint
id arxiv_https___arxiv_org_abs_2205_01970
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Non-Stationary Bandit Learning via Predictive Sampling
Liu, Yueyang
Kuang, Xu
Van Roy, Benjamin
Machine Learning
Thompson sampling has proven effective across a wide range of stationary bandit environments. However, as we demonstrate in this paper, it can perform poorly when applied to non-stationary environments. We attribute such failures to the fact that, when exploring, the algorithm does not differentiate actions based on how quickly the information acquired loses its usefulness due to non-stationarity. Building upon this insight, we propose predictive sampling, an algorithm that deprioritizes acquiring information that quickly loses usefulness. A theoretical guarantee on the performance of predictive sampling is established through a Bayesian regret bound. We provide versions of predictive sampling for which computations tractably scale to complex bandit environments of practical interest. Through numerical simulations, we demonstrate that predictive sampling outperforms Thompson sampling in all non-stationary environments examined.
title Non-Stationary Bandit Learning via Predictive Sampling
topic Machine Learning
url https://arxiv.org/abs/2205.01970