Gaussian mixture models with Wasserstein distance

By Benoit Gaujac, Ilya Feige, David Barber. 2018

9 citations recorded by the source corpus.

Generative models with both discrete and continuous latent variables are highly motivated by the structure of many real-world data sets. They present, however, subtleties in training often manifesting in the discrete latent being under leveraged. In this paper, we show that such models are more amenable to training when using the Optimal Transport framework of Wasserstein Autoencoders. We find our discrete latent variable to be fully leveraged by the model when trained, without any modifications to the objective function or significant fine tuning. Our model generates comparable samples to other approaches while using relatively simple neural networks, since the discrete latent variable carries much of the descriptive burden. Furthermore, the discrete latent provides significant control over generation.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. John Healy
2. Leland McInnes
3. Maarten Grootendorst
4. D. Angelov

In Computer Vision and Pattern Recognition

1. Leland McInnes
2. John Healy
3. Mohammad Rostami
4. D. Angelov
5. Maarten Grootendorst
6. Soheil Kolouri
7. Dmitry Kobak
8. Rahul Satija

In Probabilistic & Bayesian Methods

1. John Healy
2. Leland McInnes

Citations in the corpus

• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction (2018, methodology): “Based upon preliminary releases of a so‰ware implementation, UMAP has already found widespread use in the €elds of bioinformatics [4, 15, 37, 2, 36, 13], materials science [27, 19], and machine learning [8, 20, 17, 38] among others.”