Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation

Thomas Brouwer, Jes Frellsen, Pietro Liò

Publikation: Konferencebidrag - EJ publiceret i proceeding eller tidsskriftPaperForskningpeer review

Abstrakt

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.
OriginalsprogEngelsk
Publikationsdato9 dec. 2016
StatusUdgivet - 9 dec. 2016
BegivenhedNIPS 2016: Advances in Approximate Bayesian Inference Workshop - Room 112, Centre Convencions Internacional Barcelona, Barcelona, Spanien
Varighed: 9 dec. 20169 dec. 2016
http://approximateinference.org

Workshop

WorkshopNIPS 2016
LokationRoom 112, Centre Convencions Internacional Barcelona
Land/OmrådeSpanien
By Barcelona
Periode09/12/201609/12/2016
Internetadresse

Fingeraftryk

Dyk ned i forskningsemnerne om 'Fast Bayesian Non-Negative Matrix Factorisation and Tri-Factorisation'. Sammen danner de et unikt fingeraftryk.

Citationsformater