Interpretable ODE-style Generative Diffusion Model via Force Field Construction

Abstract

For a considerable time, researchers have focused on developing a method that establishes a deep connection between the generative diffusion model and mathematical physics. Despite previous efforts, progress has been limited to the pursuit of a single specialized method. In order to advance the interpretability of diffusion models and explore new research directions, it is essential to establish a unified ODE-style generative diffusion model. Such a model should draw inspiration from physical models and possess a clear geometric meaning. This paper aims to identify various physical models that are suitable for constructing ODE-style generative diffusion models accurately from a mathematical perspective. We then summarize these models into a unified method. Additionally, we perform a case study where we use the theoretical model identified by our method to develop a range of new diffusion model methods, and conduct experiments. Our experiments on CIFAR-10 demonstrate the effectiveness of our approach. We have constructed a computational framework that attains highly proficient results with regards to image generation speed, alongside an additional model that demonstrates exceptional performance in both Inception score and FID score. These results underscore the significance of our method in advancing the field of diffusion models.

This is empirical reproducing and experimental tries of Jianlin Su's blog use ChatGPT and Newbeing: https://kexue.fm/archives/9370 https://kexue.fm/archives/9379 https://kexue.fm/archives/9467 We make the mistakes to cite them and this paper need to be withdrawn to clearify this. We thank Jianlin reached out us and give us chance to claim