Forward Operator Estimation in Generative Models with Kernel Transfer Operators

TL;DR

This paper introduces a kernel transfer operator-based method for efficiently estimating forward operators in generative models, achieving comparable performance to existing methods with reduced computational costs.

Contribution

It proposes a novel, computationally cheaper approach for forward operator estimation in generative models using kernel transfer operators, improving efficiency and performance.

Findings

Achieves significant runtime savings compared to traditional methods.

Performs well even with small sample sizes, such as in brain imaging.

Offers competitive empirical results against strong baselines.

Abstract

Generative models which use explicit density modeling (e.g., variational autoencoders, flow-based generative models) involve finding a mapping from a known distribution, e.g. Gaussian, to the unknown input distribution. This often requires searching over a class of non-linear functions (e.g., representable by a deep neural network). While effective in practice, the associated runtime/memory costs can increase rapidly, usually as a function of the performance desired in an application. We propose a much cheaper (and simpler) strategy to estimate this mapping based on adapting known results in kernel transfer operators. We show that our formulation enables highly efficient distribution approximation and sampling, and offers surprisingly good empirical performance that compares favorably with powerful baselines, but with significant runtime savings. We show that the algorithm also performs well in small sample size settings (in brain imaging).