Clustering, factor discovery and optimal transport
arXiv:1902.10288 · doi:10.1093/imaiai/iaaa040
Abstract
The clustering problem, and more generally, latent factor discovery --or latent space inference-- is formulated in terms of the Wasserstein barycenter problem from optimal transport. The objective proposed is the maximization of the variability attributable to class, further characterized as the minimization of the variance of the Wasserstein barycenter. Existing theory, which constrains the transport maps to rigid translations, is extended to affine transformations. The resulting non-parametric clustering algorithms include k-means as a special case and exhibit more robust performance. A continuous version of these algorithms discovers continuous latent variables and generalizes principal curves. The strength of these algorithms is demonstrated by tests on both artificial and real-world data sets.
arXiv commentImproved clarity of presentation
References in corpus (4)
- Auto-Encoding Variational Bayes
- Exploiting Nonlinear Recurrence and Fractal Scaling Properties for Voice Disorder Detection
- Projection onto the probability simplex: An efficient algorithm with a simple proof, and an application
- Conditional Density Estimation, Latent Variable Discovery and Optimal Transport