Interface reliability and diffusion barrier property of Co-W barrier layer with modulated structure

Chen, Liping; Chen, Shuhui; Chen, Peixin; Cao, Liang; Li, Ming; Hang, Tao

doi:10.1016/j.matlet.2022.133501

Cited by 5 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Evaluation Metrics. We follow existing literature (Liu et al 2019a;Chou, Chen, and Ho 2023a;Chen, Song, and Li 2023;Chou, Chen, and Ho 2023b) and use the following metrics: 1) Detection Accuracy (ACC) assesses the ability of our backdoor detection method. It's the higher the better.…”

Section: Methodsmentioning

confidence: 99%

“…The model generates the target image when r is stamped on an input. To the best of our knowledge, there are three existing noiseto-image DM backdoor attacks (Chou, Chen, and Ho 2023a;Chen, Song, and Li 2023;Chou, Chen, and Ho 2023b). Their high-level goal can be illustrated in Figure 1.…”

Section: Backdoor Injection In Diffusion Modelsmentioning

confidence: 99%

“…Diffusion Models have attracted a lot of researchers, to propose different models (Ho, Jain, and Abbeel 2020;Ermon 2019, 2020;Karras et al 2022;Rombach et al 2022) and different applications (Xu et al 2022;Jeong et 2021; Popov et al 2021;Kim, Kim, and Yoon 2022;Kong et al 2021;Mei and Patel 2022;Ho et al 2022;Ruiz et al 2023). A lot of methods are proposed to deal with the slow sampling process (Song, Meng, and Ermon 2021;Lu et al 2022a,b;Zhao et al 2023;Karras et al 2022;) Though they achieve a huge success, they are vulnerable to backdoor attacks (Chou, Chen, and Ho 2023a;Chen, Song, and Li 2023;Chou, Chen, and Ho 2023b). To mitigate this issue, we propose the first backdoor detection and removal framework for diffusion models.…”

Section: Related Workmentioning

confidence: 99%

“…However, recent studies show that they are vulnerable to backdoor attacks (Chou, Chen, and Ho 2023a;Chen, Song, and Li 2023;Chou, Chen, and Ho 2023b). In traditional backdoor attacks for discriminative models such as classifiers, during training, attackers poison the training data (e.g., adding a trigger to the data and labeling them as the target class).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Elijah: Eliminating Backdoors Injected in Diffusion Models via Distribution Shift

An,

Chou,

Zhang

et al. 2024

AAAI

View full text Add to dashboard Cite

Diffusion models (DM) have become state-of-the-art generative models because of their capability of generating high-quality images from noises without adversarial training. However, they are vulnerable to backdoor attacks as reported by recent studies. When a data input (e.g., some Gaussian noise) is stamped with a trigger (e.g., a white patch), the backdoored model always generates the target image (e.g., an improper photo). However, effective defense strategies to mitigate backdoors from DMs are underexplored. To bridge this gap, we propose the first backdoor detection and removal framework for DMs. We evaluate our framework Elijah on over hundreds of DMs of 3 types including DDPM, NCSN and LDM, with 13 samplers against 3 existing backdoor attacks. Extensive experiments show that our approach can have close to 100% detection accuracy and reduce the backdoor effects to close to zero without significantly sacrificing the model utility.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Backdoor Injection In Diffusion Modelsmentioning

confidence: 99%