Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation

Thomas Brouwer; Jes Frellsen; Pietro Liò

Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation

Thomas Brouwer
, Jes Frellsen
, Pietro Liò

Algorithms

University of Cambridge

Research output: Contribution to conference - NOT published in proceeding or journal › Paper › Research › peer-review

Abstract

We present a fast variational Bayesian algorithm for performing non-negative matrix factorisation and tri-factorisation. We show that our approach achieves faster convergence per iteration and timestep (wall-clock) than Gibbs sampling and non-probabilistic approaches, and do not require additional samples to estimate the posterior. We show that in particular for matrix tri-factorisation convergence is difficult, but our variational Bayesian approach offers a fast solution, allowing the tri-factorisation approach to be used more effectively.

Original language	English
Publication date	9 Dec 2016
Publication status	Published - 9 Dec 2016
Event	NIPS 2016: Advances in Approximate Bayesian Inference Workshop - Room 112, Centre Convencions Internacional Barcelona, Barcelona, Spain Duration: 9 Dec 2016 → 9 Dec 2016 http://approximateinference.org

Workshop

Workshop	NIPS 2016
Location	Room 112, Centre Convencions Internacional Barcelona
Country/Territory	Spain
City	Barcelona
Period	09/12/2016 → 09/12/2016
Internet address	http://approximateinference.org

Keywords

variational Bayesian algorithm
non-negative matrix factorization
matrix tri-factorisation
fast convergence
posterior estimation

Access to Document

http://approximateinference.org/accepted/BrouwerEtAl2016.pdfLicence: Unspecified

Cite this

@conference{f39ae5c15fa4458bada4cc5a8f76a745,

title = "Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation",

abstract = "We present a fast variational Bayesian algorithm for performing non-negative matrix factorisation and tri-factorisation. We show that our approach achieves faster convergence per iteration and timestep (wall-clock) than Gibbs sampling and non-probabilistic approaches, and do not require additional samples to estimate the posterior. We show that in particular for matrix tri-factorisation convergence is difficult, but our variational Bayesian approach offers a fast solution, allowing the tri-factorisation approach to be used more effectively.",

keywords = "variational Bayesian algorithm, non-negative matrix factorization, matrix tri-factorisation, fast convergence, posterior estimation, variational Bayesian algorithm, non-negative matrix factorization, matrix tri-factorisation, fast convergence, posterior estimation",

author = "Thomas Brouwer and Jes Frellsen and Pietro Li{\`o}",

year = "2016",

month = dec,

day = "9",

language = "English",

note = "NIPS 2016 : Advances in Approximate Bayesian Inference Workshop ; Conference date: 09-12-2016 Through 09-12-2016",

url = "http://approximateinference.org",

}

TY - CONF

T1 - Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation

AU - Brouwer, Thomas

AU - Frellsen, Jes

AU - Liò, Pietro

PY - 2016/12/9

Y1 - 2016/12/9

N2 - We present a fast variational Bayesian algorithm for performing non-negative matrix factorisation and tri-factorisation. We show that our approach achieves faster convergence per iteration and timestep (wall-clock) than Gibbs sampling and non-probabilistic approaches, and do not require additional samples to estimate the posterior. We show that in particular for matrix tri-factorisation convergence is difficult, but our variational Bayesian approach offers a fast solution, allowing the tri-factorisation approach to be used more effectively.

AB - We present a fast variational Bayesian algorithm for performing non-negative matrix factorisation and tri-factorisation. We show that our approach achieves faster convergence per iteration and timestep (wall-clock) than Gibbs sampling and non-probabilistic approaches, and do not require additional samples to estimate the posterior. We show that in particular for matrix tri-factorisation convergence is difficult, but our variational Bayesian approach offers a fast solution, allowing the tri-factorisation approach to be used more effectively.

KW - variational Bayesian algorithm

KW - non-negative matrix factorization

KW - matrix tri-factorisation

KW - fast convergence

KW - posterior estimation

KW - variational Bayesian algorithm

KW - non-negative matrix factorization

KW - matrix tri-factorisation

KW - fast convergence

KW - posterior estimation

M3 - Paper

T2 - NIPS 2016

Y2 - 9 December 2016 through 9 December 2016

ER -

Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation

Abstract

Workshop

Keywords

Access to Document

Fingerprint

Cite this