Principled Interpolation in Normalizing Flows

10/22/2020
by   Samuel G. Fadel, et al.
0

Generative models based on normalizing flows are very successful in modeling complex data distributions using simpler ones. However, straightforward linear interpolations show unexpected side effects, as interpolation paths lie outside the area where samples are observed. This is caused by the standard choice of Gaussian base distributions and can be seen in the norms of the interpolated samples. This observation suggests that correcting the norm should generally result in better interpolations, but it is not clear how to correct the norm in an unambiguous way. In this paper, we solve this issue by enforcing a fixed norm and, hence, change the base distribution, to allow for a principled way of interpolation. Specifically, we use the Dirichlet and von Mises-Fisher base distributions. Our experimental results show superior performance in terms of bits per dimension, Fréchet Inception Distance (FID), and Kernel Inception Distance (KID) scores for interpolation, while maintaining the same generative performance.

READ FULL TEXT

page 1

page 2

page 7

page 8

page 12

page 13

research
12/24/2020

RBM-Flow and D-Flow: Invertible Flows with Discrete Energy Base Spaces

Efficient sampling of complex data distributions can be achieved using t...
research
01/29/2019

Sliced generative models

In this paper we discuss a class of AutoEncoder based generative models ...
research
12/28/2018

Consistency of Interpolation with Laplace Kernels is a High-Dimensional Phenomenon

We show that minimum-norm interpolation in the Reproducing Kernel Hilber...
research
06/18/2018

Banach Wasserstein GAN

Wasserstein Generative Adversarial Networks (WGANs) can be used to gener...
research
09/08/2023

Variations and Relaxations of Normalizing Flows

Normalizing Flows (NFs) describe a class of models that express a comple...
research
07/03/2023

Interpolation of Point Distributions for Digital Stippling

We present a new way to merge any two point distribution approaches usin...

Please sign up or login with your details

Forgot password? Click here to reset