Implementation of Gaussian Mixture Variational Autoencoder (GMVAE) for Unsupervised Clustering in PyTorch and Tensorflow. The probabilistic model is based on the model proposed by Rui Shu, which is a modification of the M2 unsupervised model proposed by Kingma et al. for semi-supervised learning.

VAE estimates latent variable（x） and sends latent variables（x） to GMM. GMM clusters latent variables（x） sent from VAE and sends mean and variance parameters of the Gaussian distribution to VAE.

本文我将介绍 VAE 针对 无监督聚类 的一个扩展： GMVAE，即基于 高斯混合模型 的VAE。 我们在 之前的文章 中已经介绍了VAE，它是一个无监督的生成模型，其良好的性能和end-to-end的性质让它在深度学习时代被广泛应用。 而GMVAE则将VAE的相关技术应用到无监督聚类问题上，其思想在于通过扩展latent variable structure提升VAE的聚类性能。 1. Introduction. 无监督聚类一直是机器学习中比较重要的题目，传统方法例如我们熟知的 k-means 或者高斯混合模型 (GMM)到 …

Nov 8, 2016 · We study a variant of the variational autoencoder model (VAE) with a Gaussian mixture as a prior distribution, with the goal of performing unsupervised clustering through deep generative models.

This project implements a powerful Variational Autoencoder (VAE) with Gaussian Mixture as a prior distribution, enabling both unsupervised clustering and high-quality image generation. The model is specifically optimized for handling custom …

Jun 12, 2018 · A variational autoencoder (VAE) provides a way of learning the probability distribution $p(x,z)$ relating an input $x$ to its latent representation $z$. In particular, the encoder $e$ maps an input $x$ to a distribution on $z$.

Oct 27, 2019 · Abstract: We propose DGG: {\textbf D}eep clustering via a {\textbf G}aussian-mixture variational autoencoder (VAE) with {\textbf G}raph embedding. To facilitate clustering, we apply Gaussian mixture model (GMM) as the prior in VAE. To handle data with complex spread, we apply graph embedding.

Mar 14, 2022 · variant of VAE with GMM as prior. goal : unsupervised clustering via DGM; Problem of regular VAE : over-regularisation \(\rightarrow\) leads to cluster degeneracy. Minimum information constraint. mitigate these problems in VAE; improve unsupervised clustering performance; 1. Introduction. Unsupervised clustering (conventional) K-means, GMM

Jan 12, 2021 · In this paper, we propose a principled unsupervised meta-learning model, namely Meta-GMVAE, based on Variational Autoencoder (VAE) and set-level variational inference.

We study a variant of the variational autoencoder model with a Gaussian mixture as a prior distribution, with the goal of performing unsupervised clustering through deep generative models.