Robust Bayesian Matrix Factorisation

Balaji Lakshminarayanan, Guillaume Bouchard, Cedric Archambeau
Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, PMLR 15:425-433, 2011.

Abstract

We analyse the noise arising in collaborative filtering when formalised as a probabilistic matrix factorisation problem. We show empirically that modelling row- and column-specific variances is important, the noise being in general non-Gaussian and heteroscedastic. We also advocate for the use of a Student-t prior for the latent features as the standard Gaussian is included as a special case. We derive several variational inference algorithms and estimate the hyperparameters by type-II maximum likelihood. Experiments on real data show that the predictive performance is significantly improved.

Cite this Paper

BibTeX
@InProceedings{pmlr-v15-lakshminarayanan11a, title = {Robust Bayesian Matrix Factorisation}, author = {Lakshminarayanan, Balaji and Bouchard, Guillaume and Archambeau, Cedric}, booktitle = {Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics}, pages = {425--433}, year = {2011}, editor = {Gordon, Geoffrey and Dunson, David and Dudík, Miroslav}, volume = {15}, series = {Proceedings of Machine Learning Research}, address = {Fort Lauderdale, FL, USA}, month = {11--13 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v15/lakshminarayanan11a/lakshminarayanan11a.pdf}, url = {https://proceedings.mlr.press/v15/lakshminarayanan11a.html}, abstract = {We analyse the noise arising in collaborative filtering when formalised as a probabilistic matrix factorisation problem. We show empirically that modelling row- and column-specific variances is important, the noise being in general non-Gaussian and heteroscedastic. We also advocate for the use of a Student-t prior for the latent features as the standard Gaussian is included as a special case. We derive several variational inference algorithms and estimate the hyperparameters by type-II maximum likelihood. Experiments on real data show that the predictive performance is significantly improved.} }
Endnote
%0 Conference Paper %T Robust Bayesian Matrix Factorisation %A Balaji Lakshminarayanan %A Guillaume Bouchard %A Cedric Archambeau %B Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2011 %E Geoffrey Gordon %E David Dunson %E Miroslav Dudík %F pmlr-v15-lakshminarayanan11a %I PMLR %P 425--433 %U https://proceedings.mlr.press/v15/lakshminarayanan11a.html %V 15 %X We analyse the noise arising in collaborative filtering when formalised as a probabilistic matrix factorisation problem. We show empirically that modelling row- and column-specific variances is important, the noise being in general non-Gaussian and heteroscedastic. We also advocate for the use of a Student-t prior for the latent features as the standard Gaussian is included as a special case. We derive several variational inference algorithms and estimate the hyperparameters by type-II maximum likelihood. Experiments on real data show that the predictive performance is significantly improved.
RIS
TY - CPAPER TI - Robust Bayesian Matrix Factorisation AU - Balaji Lakshminarayanan AU - Guillaume Bouchard AU - Cedric Archambeau BT - Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics DA - 2011/06/14 ED - Geoffrey Gordon ED - David Dunson ED - Miroslav Dudík ID - pmlr-v15-lakshminarayanan11a PB - PMLR DP - Proceedings of Machine Learning Research VL - 15 SP - 425 EP - 433 L1 - http://proceedings.mlr.press/v15/lakshminarayanan11a/lakshminarayanan11a.pdf UR - https://proceedings.mlr.press/v15/lakshminarayanan11a.html AB - We analyse the noise arising in collaborative filtering when formalised as a probabilistic matrix factorisation problem. We show empirically that modelling row- and column-specific variances is important, the noise being in general non-Gaussian and heteroscedastic. We also advocate for the use of a Student-t prior for the latent features as the standard Gaussian is included as a special case. We derive several variational inference algorithms and estimate the hyperparameters by type-II maximum likelihood. Experiments on real data show that the predictive performance is significantly improved. ER -
APA
Lakshminarayanan, B., Bouchard, G. & Archambeau, C.. (2011). Robust Bayesian Matrix Factorisation. Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 15:425-433 Available from https://proceedings.mlr.press/v15/lakshminarayanan11a.html.

Related Material