Truncated Gaussian-Mixture Variational AutoEncoder
Truncated Gaussian-Mixture Variational AutoEncoder
复制标题
DOI：
--
复制
发表时间：
2019-02
期刊：
影响因子：
--
通讯作者：
Qingyu Zhao;N. Honnorat;E. Adeli;K. Pohl
中科院分区：
文献类型：
--
作者：
Qingyu Zhao;N. Honnorat;E. Adeli;K. Pohl
变分自编码器（VAE）已经成为对低维潜在空间数据的非线性生成过程进行建模的有力工具。近年来，一些研究提出将VAE用于无监督聚类，利用混合模型捕捉潜在表征的多模态结构。然而，当存在潜在表示没有意义的异常数据样本时，这种策略是无效的，并且会污染潜在空间中关键主要聚类的估计。这个确切的问题出现在静息状态fMRI （rs-fMRI）分析的背景下，其中主要功能连接模式的聚类通常受到rs-fMRI的重噪声和许多对分析不感兴趣的次要聚类（罕见的连接模式）的阻碍。在本文中，我们提出了一种新的生成过程，其中我们使用高斯混合模型来模拟数据中的几个主要簇，并使用非信息均匀分布来捕获剩余的数据。我们将这种截断的高斯混合模型嵌入到变分自动编码器框架中，以获得一种通用的联合聚类和离群点检测方法，称为tGM-VAE。我们证明了tGM-VAE在MNIST数据集上的适用性，并在rs-fMRI连通性分析的背景下进一步验证了它。
Variation Autoencoder (VAE) has become a powerful tool in modeling the non-linear generative process of data from a low-dimensional latent space. Recently, several studies have proposed to use VAE for unsupervised clustering by using mixture models to capture the multi-modal structure of latent representations. This strategy, however, is ineffective when there are outlier data samples whose latent representations are meaningless, yet contaminating the estimation of key major clusters in the latent space. This exact problem arises in the context of resting-state fMRI (rs-fMRI) analysis, where clustering major functional connectivity patterns is often hindered by heavy noise of rs-fMRI and many minor clusters (rare connectivity patterns) of no interest to analysis. In this paper we propose a novel generative process, in which we use a Gaussian-mixture to model a few major clusters in the data, and use a non-informative uniform distribution to capture the remaining data. We embed this truncated Gaussian-Mixture model in a Variational AutoEncoder framework to obtain a general joint clustering and outlier detection approach, called tGM-VAE. We demonstrated the applicability of tGM-VAE on the MNIST dataset and further validated it in the context of rs-fMRI connectivity analysis.