On the Noise Scheduling for Generating Plausible Designs with Diffusion Models

TL;DR

This paper investigates how noise schedules in diffusion models affect the plausibility of generated structural designs, proposing new techniques and a novel schedule that significantly improve plausibility rates and image quality.

Contribution

It introduces methods to identify optimal noise levels and a new parametric noise schedule that enhances the plausibility of generated designs in diffusion models.

Findings

Plausibility depends on specific noise levels in diffusion models.

The proposed noise schedule improves plausible design generation from 83.4% to 93.5%.

FID score is reduced from 7.84 to 4.87, indicating higher quality outputs.

Abstract

Deep Generative Models (DGMs) are widely used to create innovative designs across multiple industries, ranging from fashion to the automotive sector. In addition to generating images of high visual quality, the task of structural design generation imposes more stringent constrains on the semantic expression, e.g., no floating material or missing part, which we refer to as plausibility in this work. We delve into the impact of noise schedules of diffusion models on the plausibility of the outcome: there exists a range of noise levels at which the model's performance decides the result plausibility. Also, we propose two techniques to determine such a range for a given image set and devise a novel parametric noise schedule for better plausibility. We apply this noise schedule to the training and sampling of the well-known diffusion model EDM and compare it to its default noise schedule. Compared to EDM, our schedule significantly improves the rate of plausible designs from 83.4% to 93.5% and Fr\'echet Inception Distance (FID) from 7.84 to 4.87. Further applications of advanced image editing tools demonstrate the model's solid understanding of structure.