TendiffPure: a convolutional tensor-train denoising diffusion model for purification

Abstract

Diffusion models are effective purification methods, where the noises or adversarial attacks are removed using generative approaches before pre-existing classifiers conducting classification tasks. However, the efficiency of diffusion models is still a concern, and existing solutions are based on knowledge distillation which can jeopardize the generation quality because of the small number of generation steps. Hence, we propose TendiffPure as a tensorized and compressed diffusion model for purification. Unlike the knowledge distillation methods, we directly compress U-Nets as backbones of diffusion models using tensor-train decomposition, which reduces the number of parameters and captures more spatial information in multi-dimensional data such as images. The space complexity is reduced from O(N²) to O(NR²) with R ≤ 4 as the tensor-train rank and N as the number of channels. Experimental results show that TendiffPure can more efficiently obtain high-quality purification results and outperforms the baseline purification methods on CIFAR-10, Fashion-MNIST, and MNIST datasets for two noises and one adversarial attack.

摘要

扩散模型是有效的纯化方法,在现有分类器执行分类任务之前,使用生成方法去除噪声或对抗性攻击。然而,扩散模型的效率仍然是一个问题,现有的解决方案基于知识蒸馏,由于生成步骤较少,可能会危及生成质量。因此,我们提出TendiffPure,一种用于纯化的张量化和压缩的扩散模型。与知识蒸馏方法不同,我们直接使用张量链分解压缩扩散模型的U-Net骨干网络,减少参数数量,并在多维数据(如图像)中捕获更多的空间信息。空间复杂度从O(N²)减少到O(NR²),其中R≤4为张量序列秩,N为通道数。实验结果表明,基于CIFAR-10、Fashion-MNIST和MNIST数据集,TendiffPure可以更有效地生成高质量的净化结果,并在两种噪声和一次对抗性攻击下优于基线纯化方法。