Multi-way Spectral Clustering of Augmented Multi-view Data through Deep Collective Matrix Tri-factorization

TL;DR

This paper introduces DCMTF, a novel deep learning architecture for collective matrix tri-factorization, enabling multi-way spectral clustering of diverse multi-view data to uncover latent clusters and their associations.

Contribution

It is the first deep learning-based method for collective matrix tri-factorization applicable to arbitrary multi-view data collections.

Findings

Enables multi-way spectral clustering of heterogeneous relational data.

Discovers latent clusters across multiple dimensions.

Reveals strengths of associations between clusters.

Abstract

We present the first deep learning based architecture for collective matrix tri-factorization (DCMTF) of arbitrary collections of matrices, also known as augmented multi-view data. DCMTF can be used for multi-way spectral clustering of heterogeneous collections of relational data matrices to discover latent clusters in each input matrix, across both dimensions, as well as the strengths of association across clusters. The source code for DCMTF is available on our public repository: https://bitbucket.org/cdal/dcmtf_generic