Efficient Thompson sampling for online matrix-factorization recommendation

Jaya Kawale, Hung Bui, Branislav Kveton, Long Tran Thanh, Sanjay Chawla

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

40 Citations (Scopus)

Abstract

Matrix factorization (MF) collaborative filtering is an effective and widely used method in recommendation systems. However, the problem of finding an optimal trade-off between exploration and exploitation (otherwise known as the bandit problem), a crucial problem in collaborative filtering from cold-start, has not been previously addressed. In this paper, we present a novel algorithm for online MF recommendation that automatically combines finding the most relevant items with exploring new or less-recommended items. Our approach, called Particle Thompson sampling for MF (PTS), is based on the general Thompson sampling framework, but augmented with a novel efficient online Bayesian probabilistic matrix factorization method based on the Rao-Blackwellized particle filter. Extensive experiments in collaborative filtering using several real-world datasets demonstrate that PTS significantly outperforms the current state-of-the-arts.

Original languageEnglish
Title of host publicationAdvances in Neural Information Processing Systems
PublisherNeural information processing systems foundation
Pages1297-1305
Number of pages9
Volume2015-January
Publication statusPublished - 2015
Externally publishedYes
Event29th Annual Conference on Neural Information Processing Systems, NIPS 2015 - Montreal, Canada
Duration: 7 Dec 201512 Dec 2015

Other

Other29th Annual Conference on Neural Information Processing Systems, NIPS 2015
CountryCanada
CityMontreal
Period7/12/1512/12/15

Fingerprint

Factorization
Collaborative filtering
Sampling
Recommender systems
Experiments

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Signal Processing

Cite this

Kawale, J., Bui, H., Kveton, B., Thanh, L. T., & Chawla, S. (2015). Efficient Thompson sampling for online matrix-factorization recommendation. In Advances in Neural Information Processing Systems (Vol. 2015-January, pp. 1297-1305). Neural information processing systems foundation.

Efficient Thompson sampling for online matrix-factorization recommendation. / Kawale, Jaya; Bui, Hung; Kveton, Branislav; Thanh, Long Tran; Chawla, Sanjay.

Advances in Neural Information Processing Systems. Vol. 2015-January Neural information processing systems foundation, 2015. p. 1297-1305.

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

Kawale, J, Bui, H, Kveton, B, Thanh, LT & Chawla, S 2015, Efficient Thompson sampling for online matrix-factorization recommendation. in Advances in Neural Information Processing Systems. vol. 2015-January, Neural information processing systems foundation, pp. 1297-1305, 29th Annual Conference on Neural Information Processing Systems, NIPS 2015, Montreal, Canada, 7/12/15.
Kawale J, Bui H, Kveton B, Thanh LT, Chawla S. Efficient Thompson sampling for online matrix-factorization recommendation. In Advances in Neural Information Processing Systems. Vol. 2015-January. Neural information processing systems foundation. 2015. p. 1297-1305
Kawale, Jaya ; Bui, Hung ; Kveton, Branislav ; Thanh, Long Tran ; Chawla, Sanjay. / Efficient Thompson sampling for online matrix-factorization recommendation. Advances in Neural Information Processing Systems. Vol. 2015-January Neural information processing systems foundation, 2015. pp. 1297-1305
@inproceedings{91434477b75c406a9febf5a48a32e170,
title = "Efficient Thompson sampling for online matrix-factorization recommendation",
abstract = "Matrix factorization (MF) collaborative filtering is an effective and widely used method in recommendation systems. However, the problem of finding an optimal trade-off between exploration and exploitation (otherwise known as the bandit problem), a crucial problem in collaborative filtering from cold-start, has not been previously addressed. In this paper, we present a novel algorithm for online MF recommendation that automatically combines finding the most relevant items with exploring new or less-recommended items. Our approach, called Particle Thompson sampling for MF (PTS), is based on the general Thompson sampling framework, but augmented with a novel efficient online Bayesian probabilistic matrix factorization method based on the Rao-Blackwellized particle filter. Extensive experiments in collaborative filtering using several real-world datasets demonstrate that PTS significantly outperforms the current state-of-the-arts.",
author = "Jaya Kawale and Hung Bui and Branislav Kveton and Thanh, {Long Tran} and Sanjay Chawla",
year = "2015",
language = "English",
volume = "2015-January",
pages = "1297--1305",
booktitle = "Advances in Neural Information Processing Systems",
publisher = "Neural information processing systems foundation",

}

TY - GEN

T1 - Efficient Thompson sampling for online matrix-factorization recommendation

AU - Kawale, Jaya

AU - Bui, Hung

AU - Kveton, Branislav

AU - Thanh, Long Tran

AU - Chawla, Sanjay

PY - 2015

Y1 - 2015

N2 - Matrix factorization (MF) collaborative filtering is an effective and widely used method in recommendation systems. However, the problem of finding an optimal trade-off between exploration and exploitation (otherwise known as the bandit problem), a crucial problem in collaborative filtering from cold-start, has not been previously addressed. In this paper, we present a novel algorithm for online MF recommendation that automatically combines finding the most relevant items with exploring new or less-recommended items. Our approach, called Particle Thompson sampling for MF (PTS), is based on the general Thompson sampling framework, but augmented with a novel efficient online Bayesian probabilistic matrix factorization method based on the Rao-Blackwellized particle filter. Extensive experiments in collaborative filtering using several real-world datasets demonstrate that PTS significantly outperforms the current state-of-the-arts.

AB - Matrix factorization (MF) collaborative filtering is an effective and widely used method in recommendation systems. However, the problem of finding an optimal trade-off between exploration and exploitation (otherwise known as the bandit problem), a crucial problem in collaborative filtering from cold-start, has not been previously addressed. In this paper, we present a novel algorithm for online MF recommendation that automatically combines finding the most relevant items with exploring new or less-recommended items. Our approach, called Particle Thompson sampling for MF (PTS), is based on the general Thompson sampling framework, but augmented with a novel efficient online Bayesian probabilistic matrix factorization method based on the Rao-Blackwellized particle filter. Extensive experiments in collaborative filtering using several real-world datasets demonstrate that PTS significantly outperforms the current state-of-the-arts.

UR - http://www.scopus.com/inward/record.url?scp=84965114630&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84965114630&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84965114630

VL - 2015-January

SP - 1297

EP - 1305

BT - Advances in Neural Information Processing Systems

PB - Neural information processing systems foundation

ER -