Efficient Thompson sampling for online matrix-factorization recommendation
Efficient Thompson sampling for online matrix-factorization recommendation
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 relevantitems with exploring new or less-recommended items.Our approach, called Particle Thompson Sampling for Matrix-Factorization, 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 our proposed algorithm significantly outperforms the current state-of-the-arts.
1297-1305
Association for Computing Machinery
Kawale, Jaya
679ace39-ef5d-4ed2-9be6-adca97cd402c
Bui, Hung
eeba59db-466b-4278-bf57-7ac13883be5f
Kveton, Branislav
416796f9-3f66-4d2c-98a6-3575e9102775
Tran-Thanh, Long
e0666669-d34b-460e-950d-e8b139fab16c
Chawla, Sanjay
bc56b9e4-fa0c-4a03-8565-4506bc5a42e4
10 December 2015
Kawale, Jaya
679ace39-ef5d-4ed2-9be6-adca97cd402c
Bui, Hung
eeba59db-466b-4278-bf57-7ac13883be5f
Kveton, Branislav
416796f9-3f66-4d2c-98a6-3575e9102775
Tran-Thanh, Long
e0666669-d34b-460e-950d-e8b139fab16c
Chawla, Sanjay
bc56b9e4-fa0c-4a03-8565-4506bc5a42e4
Kawale, Jaya, Bui, Hung, Kveton, Branislav, Tran-Thanh, Long and Chawla, Sanjay
(2015)
Efficient Thompson sampling for online matrix-factorization recommendation.
In NIPS'15: Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 1.
vol. 1,
Association for Computing Machinery.
.
Record type:
Conference or Workshop Item
(Paper)
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 relevantitems with exploring new or less-recommended items.Our approach, called Particle Thompson Sampling for Matrix-Factorization, 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 our proposed algorithm significantly outperforms the current state-of-the-arts.
Text
paper.pdf
- Version of Record
Restricted to Repository staff only
Request a copy
More information
Accepted/In Press date: 10 September 2015
e-pub ahead of print date: 10 December 2015
Published date: 10 December 2015
Venue - Dates:
Neural Information Processing Systems (NIPS 2015), , Montreal, Canada, 2015-12-07 - 2015-12-12
Organisations:
Agents, Interactions & Complexity
Identifiers
Local EPrints ID: 387963
URI: http://eprints.soton.ac.uk/id/eprint/387963
ISSN: 1049-5258
PURE UUID: 9955c551-38e9-4ad1-83ae-7c81ee6d09cb
Catalogue record
Date deposited: 17 Feb 2016 13:45
Last modified: 16 Mar 2024 05:47
Export record
Contributors
Author:
Jaya Kawale
Author:
Hung Bui
Author:
Branislav Kveton
Author:
Long Tran-Thanh
Author:
Sanjay Chawla
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics