Phase Transition Structure of Variational Bayesian Nonnegative Matrix Factorization

Phase Transition Structure of Variational Bayesian Nonnegative Matrix Factorization
复制标题

变分贝叶斯非负矩阵分解的相变结构

DOI：
10.1007/978-3-319-68612-7_17
复制
发表时间：
2017
期刊：
Proc. of ICANN
影响因子：
--
通讯作者：
Watanabe Sumio
Watanabe Sumio
中科院分区：
--
文献类型：
--
作者：
Kohjima Masahiro;Watanabe Sumio

文献摘要

相似文献

本文从理论上阐明了变分贝叶斯非负矩阵分解（VBNMF）的相变结构。通过对变分推理中目标泛函的渐近分析，我们发现超参数极大地改变了VBNMF的变分后验分布，我们称这种现象为VBNMF的相变.我们还讨论了数值实验证明我们的理论结果。
In this paper, we theoretically clarify the phase transition structure of the variational Bayesian nonnegative matrix factorization (VBNMF). By asymptotic analysis of the objective functional in variational inference, we find that the variational posterior distribution of the VBNMF is drastically changed by hyperparameters; we call this phenomenonphase transitionof the VBNMF. We also discuss a numerical experiment demonstrating our theoretical results.