Abstract:A recent research area in unsupervised learning is the combination of representation learning with deep neural networks and data clustering. The success of deep learning for supervised tasks is widely established. However, recent research has demonstrated how neural networks are able to learn representations to improve clustering in their intermediate feature space, using specific regularizations. By considering representation learning and clustering as a joint task, models learn clustering-friendly spaces and outperform two-stage approaches where dimensionality reduction and clustering are performed separately. Recently, this idea has been extended to topology-preserving clustering models, known as self-organizing maps (SOM). This work is a thorough study on the deep embedded self-organizing map (DESOM), a model composed of an autoencoder and a SOM layer, training jointly the code vectors and network weights to learn SOM-friendly representations. In other words, SOM induces a form a regularization to improve the quality of quantization and topology in latent space. After detailing the architecture, loss and training algorithm, we study hyperparameters with a series of experiments. Different SOM-based models are evaluated in terms of clustering, visualization and classification on benchmark datasets. We study benefits and trade-offs of joint representation learning and self-organization. DESOM achieves competitive results, requires no pretraining and produces topologically organized visualizations.

Non-Euclidean Self-Organizing Maps

A New Adaptive Self-Organizing Map.

Self organising maps for visualising and modelling

Neural maps in remote sensing image analysis

Deep embedded self-organizing maps for joint representation learning and topology-preserving clustering

A Study of Parallel Self-Organizing Map

Deep Self-Organizing Map for visual classification

Self-organizing Map As a New Method for Clustering and Data Analysis

Clustering High-Dimensional Data Using Growing Som

Visualizing Clusters Based on the Serf- Organizing Map

Dynamic and Adaptive Self Organizing Maps Applied to High Dimensional Large Scale Text Clustering

A Rigorous Link Between Self-Organizing Maps and Gaussian Mixture Models

Self-organizing subspace clustering for high-dimensional and multi-view data

An Expanding Self-Organizing Neural Network for the Traveling Salesman Problem

Kernel-Based Self-Organizing Map Clustering

Neuro-mimetic Task-free Unsupervised Online Learning with Continual Self-Organizing Maps

Self-Organizing Maps for Exploration of Partially Observed Data and Imputation of Missing Values

ConSOM: A conceptional self-organizing map model for text clustering

Improving Self-Organizing Maps with Unsupervised Feature Extraction

Fast Algorithm and Implementation of Dissimilarity Self-Organizing Maps