Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

16

ICML
2010
IEEE

favoriteEmaildiscussreport

227views Machine Learning» more ICML 2010»

Implicit Regularization in Variational Bayesian Matrix Factorization

13 years 2 months ago

Implicit Regularization in Variational Bayesian Matrix Factorization

Download www.icml2010.org

Matrix factorization into the product of lowrank matrices induces non-identifiability, i.e., the mapping between the target matrix and factorized matrices is not one-to-one. In this paper, we theoretically investigate the influence of non-identifiability on Bayesian matrix factorization. More specifically, we show that a variational Bayesian method involves regularization effect even when the prior is non-informative, which is intrinsically different from the maximum a posteriori approach. We also extend our analysis to empirical Bayes scenarios where hyperparameters are also learned from data.

Shinichi Nakajima, Masashi Sugiyama

Real-time Traffic

Bayesian Matrix Factorization | ICML 2010 | Machine Learning | Matrix Factorization | Variational Bayesian Method |

claim paper

Related Content

» Analysis of Variational Bayesian Matrix Factorization

» Fast Variational Bayesian Inference for NonConjugate Matrix Factorization Models

» Bayesian probabilistic matrix factorization using Markov chain Monte Carlo

» Lowrank matrix completion by variational sparse Bayesian learning

» Bayesian Matrix Factorization with Side Information and Dirichlet Process Mixtures

» Bayesian Nonparametric Matrix Factorization for Recorded Music

» A sparse variational Bayesian approach for fMRI data analysis

» Hierarchical multilinear models for multiway data

» Dynamic Exponential Family Matrix Factorization

Post Info
More Details (n/a)

Added	12 Feb 2011
Updated	12 Feb 2011
Type	Journal
Year	2010
Where	ICML
Authors	Shinichi Nakajima, Masashi Sugiyama

Comments (0)