Variational Deep Embedding: An Unsupervised and Generative Approach to Clustering

Abstract

Clustering is among the most fundamental tasks in computer vision and machine learning. In this paper, we propose Variational Deep Embedding (VaDE), a novel unsupervised generative clustering approach within the framework of Variational Auto-Encoder (VAE). Specifically, VaDE models the data generative procedure with a Gaussian Mixture Model (GMM) and a deep neural network (DNN): 1) the GMM picks a cluster; 2) from which a latent embedding is generated; 3) then the DNN decodes the latent embedding into observables. Inference in VaDE is done in a variational way: a different DNN is used to encode observables to latent embeddings, so that the evidence lower bound (ELBO) can be optimized using Stochastic Gradient Variational Bayes (SGVB) estimator and the reparameterization trick. Quantitative comparisons with strong baselines are included in this paper, and experimental results show that VaDE significantly outperforms the state-of-the-art clustering methods on 4 benchmarks from various modalities. Moreover, by VaDE's generative nature, we show its capability of generating highly realistic samples for any specified cluster, without using supervised information during training. Lastly, VaDE is a flexible and extensible framework for unsupervised generative clustering, more general mixture models than GMM can be easily plugged in.

8 pages, 6 figures, accepted by IJCAI2017

References in corpus (7)

• Conditional Generative Adversarial Nets
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• Improved Techniques for Training GANs
• DRAW: A Recurrent Neural Network For Image Generation
• Deep Convolutional Inverse Graphics Network
• Plug & Play Generative Networks: Conditional Iterative Generation of Images in Latent Space
• Infinite Variational Autoencoder for Semi-Supervised Learning

Cited by in corpus (35)

• Learning Discrete Representations via Information Maximizing Self-Augmented Training
• Clustering with Deep Learning: Taxonomy and New Methods
• SpectralNet: Spectral Clustering using Deep Neural Networks
• Topic-Guided Variational Autoencoders for Text Generation
• AstroVaDEr: Astronomical Variational Deep Embedder for Unsupervised Morphological Classification of Galaxies and Synthetic Image Generation
• Deblending galaxies with Variational Autoencoders: a joint multi-band, multi-instrument approach
• Deep Discriminative Clustering Analysis
• Deep Unsupervised Clustering Using Mixture of Autoencoders
• GAP: Generalizable Approximate Graph Partitioning Framework
• Unsupervised Semantic Aggregation and Deformable Template Matching for Semi-Supervised Learning
• Gaussian Process Prior Variational Autoencoders
• RDEC: Integrating Regularization into Deep Embedded Clustering for Imbalanced Datasets
• Latent space conditioning for improved classification and anomaly detection
• Competitive Training of Mixtures of Independent Deep Generative Models
• Deep Discriminative Latent Space for Clustering
• Music FaderNets: Controllable Music Generation Based On High-Level Features via Low-Level Feature Modelling
• A Framework for Deep Constrained Clustering -- Algorithms and Advances
• Supporting Clustering with Contrastive Learning
• Deep Image Clustering with Category-Style Representation
• Improving Deep Image Clustering With Spatial Transformer Layers
• Consistent Representation Learning for High Dimensional Data Analysis
• Deep Variational Transfer: Transfer Learning through Semi-supervised Deep Generative Models
• Generative Modelling for Controllable Audio Synthesis of Expressive Piano Performance
• Dissimilarity Mixture Autoencoder for Deep Clustering
• Unsupervised Star Galaxy Classification with Cascade Variational Auto-Encoder
• A Framework for Joint Unsupervised Learning of Cluster-Aware Embedding for Heterogeneous Networks
• Deep Markov Spatio-Temporal Factorization
• Document Hashing with Mixture-Prior Generative Models
• Adversarially Learned Mixture Model
• Unsupervised Representation Adversarial Learning Network: from Reconstruction to Generation
• Chaotic-to-Fine Clustering for Unlabeled Plant Disease Images
• Reducing Conservativeness Oriented Offline Reinforcement Learning
• Deep Unsupervised Clustering with Clustered Generator Model
• Adversarial confidence and smoothness regularizations for scalable unsupervised discriminative learning
• Deep Clustering with Measure Propagation