Hybrid Generative Models for Two-Dimensional Datasets

TL;DR

This paper introduces a novel, basis-space approach for generative modeling of two-dimensional datasets, effectively capturing correlations and extending to scientific data, with comprehensive comparisons and a new evaluation metric.

Contribution

It proposes a general basis-space generative framework for 2D data, applicable across domains, and introduces a new metric for evaluating generation quality beyond pixel space.

Findings

The basis-space approach improves correlation modeling in generated data.

The method is effective for both imaging and scientific datasets.

A new evaluation metric better captures generation deficiencies.

Abstract

Two-dimensional array-based datasets are pervasive in a variety of domains. Current approaches for generative modeling have typically been limited to conventional image datasets and performed in the pixel domain which do not explicitly capture the correlation between pixels. Additionally, these approaches do not extend to scientific and other applications where each element value is continuous and is not limited to a fixed range. In this paper, we propose a novel approach for generating two-dimensional datasets by moving the computations to the space of representation bases and show its usefulness for two different datasets, one from imaging and another from scientific computing. The proposed approach is general and can be applied to any dataset, representation basis, or generative model. We provide a comprehensive performance comparison of various combinations of generative models and representation basis spaces. We also propose a new evaluation metric which captures the deficiency of generating images in pixel space.