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。
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Carroll JD, Chang JJ, 1970. Analysis of individual differences in multidimensional scaling via an N-way generalization of “Eckart-Young” decomposition. Psychometrika, 35(3):283–319. https://doi.org/10.1007/BF02310791
Croce F, Hein M, 2020. Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks. Proc 37th Int Conf on Machine Learning, Article 206.
Dhariwal P, Nichol A, 2021. Diffusion models beat GANs on image synthesis. Proc 35th Conf on Neural Information Processing Systems, p.8780–8794.
Gao Q, Li ZL, Zhang JP, et al., 2023. CoreDiff: contextual error-modulated generalized diffusion model for low-dose CT denoising and generalization. IEEE Trans Med Imag, early access. https://doi.org/10.1109/TMI.2023.3320812
Giovannetti V, Montangero S, Fazio R, 2008. Quantum multiscale entanglement renormalization ansatz channels. Phys Rev Lett, 101(18):180503. https://doi.org/10.1103/PhysRevLett.101.180503
Hitchcock FL, 1927. The expression of a tensor or a polyadic as a sum of products. J Math Phys, 6(1–4):164–189. https://doi.org/10.1002/sapm192761164
Ho J, Salimans T, 2021. Classifier-free diffusion guidance. Proc Workshop on Deep Generative Models and Downstream Applications.
Ho J, Jain A, Abbeel P, 2020. Denoising diffusion probabilistic models. Proc 34th Int Conf on Neural Information Processing Systems, Article 574.
Hu EJ, Shen YL, Wallis P, et al., 2022. LoRA: low-rank adaptation of large language models. Proc 10th Int Conf on Learning Representations.
Krizhevsky A, Hinton G, 2009. Learning Multiple Layers of Features from Tiny Images. Technical Report. University of Toronto, Toronto, Canada.
LeCun Y, Bottou L, Bengio Y, et al., 1998. Gradient-based learning applied to document recognition. Proc IEEE, 86(11):2278–2324. https://doi.org/10.1109/5.726791
Li C, Sun Z, Yu JS, et al., 2019. Low-rank embedding of kernels in convolutional neural networks under random shuffling. Proc IEEE Int Conf on Acoustics, p.3022–3026. https://doi.org/10.1109/ICASSP.2019.8682265
Luo YS, Zhao XL, Meng DY, et al., 2022. HLRTF: hierarchical low-rank tensor factorization for inverse problems in multi-dimensional imaging. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.19281–19290. https://doi.org/10.1109/CVPR52688.2022.01870
Meng CL, Rombach R, Gao RQ, et al., 2023. On distillation of guided diffusion models. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.14297–14306. https://doi.org/10.1109/CVPR52729.2023.01374
Nichol AQ, Dhariwal P, 2021. Improved denoising diffusion probabilistic models. Proc 38th Int Conf on Machine Learning, p.8162–8171.
Nie WL, Guo B, Huang YJ, et al., 2022. Diffusion models for adversarial purification. Proc 39th Int Conf on Machine Learning, p.16805–16827.
Oseledets IV, 2011. Tensor-train decomposition. SIAM J Sci Comput, 33(5):2295–2317. https://doi.org/10.1137/090752286
Ronneberger O, Fischer P, Brox T, 2015. U-Net: convolutional networks for biomedical image segmentation. Proc 18th Int Conf on Medical Image Computing and Computer-Assisted Intervention, p.234–241. https://doi.org/10.1007/978-3-319-24574-4_28
Song JM, Meng CL, Ermon S, 2021. Denoising diffusion implicit models. Proc 9th Int Conf on Learning Representations.
Song Y, Garg S, Shi JX, et al., 2020. Sliced score matching: a scalable approach to density and score estimation. Proc 35th Uncertainty in Artificial Intelligence Conf, p.574–584.
Song Y, Dhariwal P, Chen M, et al., 2023. Consistency models. Proc 40th Int Conf on Machine Learning, Article 1335.
Su JH, Byeon W, Kossaifi J, et al., 2020. Convolutional tensor-train LSTM for spatio-temporal learning. Proc 34th Int Conf on Neural Information Processing Systems, Article 1150.
Tucker LR, 1966. Some mathematical notes on three-mode factor analysis. Psychometrika, 31(3):279–311. https://doi.org/10.1007/BF02289464
Vahdat A, Kreis K, Kautz J, 2021. Score-based generative modeling in latent space. Proc 35th Conf on Neural Information Processing Systems.
Vincent P, 2011. A connection between score matching and denoising autoencoders. Neur Comput, 23(7):1661–1674. https://doi.org/10.1162/NECO_a_00142
Xiao H, Rasul K, Vollgraf R, 2017. Fashion-MNIST: a novel image dataset for benchmarking machine learning algorithms. https://arxiv.org/abs/1708.07747
Zhao QB, Zhou GX, Xie SL, et al., 2016. Tensor ring decomposition. https://arxiv.org/abs/1606.05535
Author information
Authors and Affiliations
Contributions
Mingyuan BAI designed the research. Derun ZHOU processed the data. Mingyuan BAI drafted the paper. Qibin ZHAO helped organize the paper. Mingyuan BAI and Derun ZHOU revised and finalized the paper.
Corresponding author
Ethics declarations
All the authors declare that they have no conflict of interest.
Rights and permissions
About this article
Cite this article
Bai, M., Zhou, D. & Zhao, Q. TendiffPure: a convolutional tensor-train denoising diffusion model for purification. Front Inform Technol Electron Eng 25, 160–169 (2024). https://doi.org/10.1631/FITEE.2300392
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/FITEE.2300392