Bootstrapping Diffusion: Diffusion Model Training Leveraging Partial and Corrupted Data

TL;DR

This paper explores training diffusion models using partial and corrupted data, proposing a residual score function approach with theoretical guarantees for improved data efficiency and generalization.

Contribution

It introduces a novel method for training diffusion models with partial data views, supported by theoretical analysis and error bounds.

Findings

The residual score function approach reduces generalization error.

Training separate models per view improves data utilization.

The method achieves near first-order optimal data efficiency.

Abstract

Training diffusion models requires large datasets. However, acquiring large volumes of high-quality data can be challenging, for example, collecting large numbers of high-resolution images and long videos. On the other hand, there are many complementary data that are usually considered corrupted or partial, such as low-resolution images and short videos. Other examples of corrupted data include videos that contain subtitles, watermarks, and logos. In this study, we investigate the theoretical problem of whether the above partial data can be utilized to train conventional diffusion models. Motivated by our theoretical analysis in this study, we propose a straightforward approach of training diffusion models utilizing partial data views, where we consider each form of complementary data as a view of conventional data. Our proposed approach first trains one separate diffusion model for each individual view, and then trains a model for predicting the residual score function. We prove generalization error bounds, which show that the proposed diffusion model training approach can achieve lower generalization errors if proper regularizations are adopted in the residual score function training. In particular, we prove that the difficulty in training the residual score function scales proportionally with the signal correlations not captured by partial data views. Consequently, the proposed approach achieves near first-order optimal data efficiency.