arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2303.04772cs.LGcs.CVmath.PRstat.ML
收藏

多层扩散:用于图像生成的无限维基于分数的扩散模型

Multilevel Diffusion: Infinite Dimensional Score-Based Diffusion Models for Image Generation

TU Berlin(柏林工业大学);Emory University(埃默里大学)

Paul Hagemann, Sophie Mildenberger, Lars Ruthotto, Gabriele Steidl, Nicole Tianjiao Yang

AI总结:

本文将基于分数的扩散模型推广到无限维函数空间,通过迹类算子和算子网络实现跨分辨率的多层训练,并在多个数据集上验证了其优势。

摘要:

基于分数的扩散模型(SBDM)近年来已成为图像生成的最先进方法。现有的 SBDM 通常在有限维设置中构建,其中图像被视为有限大小的张量。本文在无限维设置中发展 SBDM,即我们将训练数据建模为支撑在矩形域上的函数。除了追求以更高分辨率生成图像之外,我们的主要动机是创建一个适定的无限维学习问题,使其能够在多个分辨率层级上一致地离散化。我们由此旨在获得能够跨不同分辨率层级泛化的扩散模型,并提高训练过程的效率。我们展示了如何在无限维设置中克服当前 SBDM 方法的两个不足。首先,我们使用迹类算子修改前向过程,以确保潜在分布在无限维设置中是良定义的,并为有限维近似推导反向过程。其次,我们说明用算子网络近似分数函数对多层训练是有益的。在推导了离散化的收敛性以及多层训练的近似之后,我们在一个合成高斯混合示例、MNIST 数据集以及由非线性二维反应扩散方程生成的数据集上展示了我们的无限维 SBDM 方法的一些实际优势。

英文摘要:

Score-based diffusion models (SBDM) have recently emerged as state-of-the-art approaches for image generation. Existing SBDMs are typically formulated in a finite-dimensional setting, where images are considered as tensors of finite size. This paper develops SBDMs in the infinite-dimensional setting, that is, we model the training data as functions supported on a rectangular domain. In addition to the quest for generating images at ever-higher resolutions, our primary motivation is to create a well-posed infinite-dimensional learning problem that we can discretize consistently on multiple resolution levels. We thereby intend to obtain diffusion models that generalize across different resolution levels and improve the efficiency of the training process. We demonstrate how to overcome two shortcomings of current SBDM approaches in the infinite-dimensional setting. First, we modify the forward process using trace class operators to ensure that the latent distribution is well-defined in the infinite-dimensional setting and derive the reverse processes for finite-dimensional approximations. Second, we illustrate that approximating the score function with an operator network is beneficial for multilevel training. After deriving the convergence of the discretization and the approximation of multilevel training, we demonstrate some practical benefits of our infinite-dimensional SBDM approach on a synthetic Gaussian mixture example, the MNIST dataset, and a dataset generated from a nonlinear 2D reaction-diffusion equation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑