Improving Particle Thompson Sampling through Regenerative Particles

Zeyu Zhou, Bruce Hajek

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper proposes regenerative particle Thompson sampling (RPTS) as an improvement of particle Thompson sampling (PTS) for solving general stochastic bandit problems. PTS approximates Thompson sampling by replacing the continuous posterior distribution with a discrete distribution supported at a set of weighted static particles. PTS is flexible but may suffer from poor performance due to the tendency of the probability mass to concentrate on a small number of particles. RPTS exploits the particle weight dynamics of PTS and uses non-static particles: it deletes a particle if its probability mass gets sufficiently small and regenerates new particles in the vicinity of the surviving particles. Empirical evidence shows uniform improvement across a set of representative bandit problems without increasing the number of particles.

Original languageEnglish (US)
Title of host publication2023 57th Annual Conference on Information Sciences and Systems, CISS 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665451819
DOIs
StatePublished - 2023
Externally publishedYes
Event57th Annual Conference on Information Sciences and Systems, CISS 2023 - Baltimore, United States
Duration: Mar 22 2023Mar 24 2023

Publication series

Name2023 57th Annual Conference on Information Sciences and Systems, CISS 2023

Conference

Conference57th Annual Conference on Information Sciences and Systems, CISS 2023
Country/TerritoryUnited States
CityBaltimore
Period3/22/233/24/23

Keywords

  • Thompson sampling
  • particles
  • stochastic bandit

ASJC Scopus subject areas

  • Computer Science Applications
  • Hardware and Architecture
  • Information Systems
  • Artificial Intelligence
  • Information Systems and Management
  • Safety, Risk, Reliability and Quality

Fingerprint

Dive into the research topics of 'Improving Particle Thompson Sampling through Regenerative Particles'. Together they form a unique fingerprint.

Cite this