While convolutional neural networks (CNNs) have achieved success in computer vision tasks, it is vulnerable to backdoor attacks. Such attacks could mislead the victim model to make attacker-chosen prediction with a specific trigger pattern. Until now, the trigger injection of existing attacks is mainly limited to spatial domain. Recent works take advantage of perceptual properties of planting specific patterns in the frequency domain, which only reflect indistinguishable pixel-wise perturbations in pixel domain. However, in the black-box setup, the inaccessibility of training process often renders more complex trigger designs. Existing frequency attacks simply handcraft the magnitude of spectrum, introducing anomaly frequency disparities between clean and poisoned data and taking risks of being removed by image processing operations (such as lossy compression and filtering). In this paper, we propose a robust low-frequency black-box backdoor attack (LFBA), which minimally perturbs low-frequency components of frequency spectrum and maintains the perceptual similarity in spatial space simultaneously. The key insight of our attack restrict the search for the optimal trigger to low-frequency region that can achieve high attack effectiveness, robustness against image transformation defenses and stealthiness in dual space. We utilize simulated annealing (SA), a form of evolutionary algorithm, to optimize the properties of frequency trigger including the number of manipulated frequency bands and the perturbation of each frequency component, without relying on the knowledge from the victim classifier. Extensive experiments on real-world datasets verify the effectiveness and robustness of LFBA against image processing operations and the state-of-the-art backdoor defenses, as well as its inherent stealthiness in both spatial and frequency space, making it resilient against frequency inspection.
MakeupAttack: Feature Space Black-box Backdoor Attack on Face Recognition via Makeup Transfer
Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络空间安全学院）
Tsinghua University（清华大学）；RealAI（瑞莱智慧）；Institute for AI（人工智能研究院）；BNRist Center（北京国家信息科学技术研究中心）；Tsinghua-Bosch Joint ML Center（清华-博世联合机器学习中心）；THBI Lab（清华-北京生物医学工程研究院实验室）；Dept. of Comp. Sci. and Tech.（计算机科学与技术系）
Although deep neural networks (DNNs) have made rapid progress in recent years, they are vulnerable in adversarial environments. A malicious backdoor could be embedded in a model by poisoning the training dataset, whose intention is to make the infected model give wrong predictions during inference when the specific trigger appears. To mitigate the potential threats of backdoor attacks, various backdoor detection and defense methods have been proposed. However, the existing techniques usually require the poisoned training data or access to the white-box model, which is commonly unavailable in p…
BlackMirror: Black-Box Backdoor Detection for Text-to-Image Models via Instruction-Response Deviation
Institute of Information Engineering, CAS（中国科学院信息工程研究所）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络安全学院）；State Key Laboratory of AI Safety, Institute of Computing Technology, CAS（中国科学院计算技术研究所人工智能安全国家重点实验室）；School of Computer Science and Technology, University of Chinese Academy of Sciences（中国科学院大学计算机科学与技术学院）；BDKM, University of Chinese Academy of Sciences（中国科学院大学BDKM）；School of Computer Science and Technology, Beijing Institute of Technology（北京理工大学计算机科学与技术学院）；School of Cyber Science and Tech., Shenzhen Campus, Sun Yat-sen University（中山大学深圳校区网络安全科学与技术学院）
Nanyang Technological University（南洋理工大学）；City University of Hong Kong（香港城市大学）
Deep neural networks (DNN) have shown great success in many computer vision applications. However, they are also known to be susceptible to backdoor attacks. When conducting backdoor attacks, most of the existing approaches assume that the targeted DNN is always available, and an attacker can always inject a specific pattern to the training data to further fine-tune the DNN model. However, in practice, such attack may not be feasible as the DNN model is encrypted and only available to the secure enclave. In this paper, we propose a novel black-box backdoor attack technique on face recognition …
SenseTime Research（商汤科技研究院）；Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；Beihang University（北京航空航天大学）
Extensive evidence has demonstrated that deep neural networks (DNNs) are vulnerable to backdoor attacks, which motivates the development of backdoor attacks detection. Most detection methods are designed to verify whether a model is infected with presumed types of backdoor attacks, yet the adversary is likely to generate diverse backdoor attacks in practice that are unforeseen to defenders, which challenge current detection strategies. In this paper, we focus on this more challenging scenario and propose a universal backdoor attacks detection method named Adaptive Adversarial Probe (A2P). Spec…
PersGuard: Preventing Malicious Personalization in Text-to-Image Diffusion Models via Model Backdoors
Institute of Information Engineering, Chinese Academy of Sciences（信息工程研究所，中国科学院）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络安全学院）；BraneMatrix AI；School of Cyber Science and Technology, Shenzhen Campus, Sun Yat-sen University（中山大学深圳校区计算机科学与技术学院）
Watermarking has become a plausible candidate for ownership verification and intellectual property protection of deep neural networks. Regarding image classification neural networks, current watermarking schemes uniformly resort to backdoor triggers. However, injecting a backdoor into a neural network requires knowledge of the training dataset, which is usually unavailable in the real-world commercialization. Meanwhile, established watermarking schemes oversight the potential damage of exposed evidence during ownership verification and the watermarking algorithms themselves. Those concerns dec…
ODPure: Backdoor Purification for Object Detection via Ensemble Corruption Consensus
Changsha University of Science & Technology（长沙理工大学）；Xiangtan University（湘潭大学）；Yunnan University（云南大学）；Huazhong University of Science and Technology（华中科技大学）；City University of Hong Kong（香港城市大学）；Griffith University（格里菲斯大学）
Anhui University（安徽大学）；Anqing Normal University（安庆师范大学）
Fooling deep neural networks (DNNs) with the black-box optimization has become a popular adversarial attack fashion, as the structural prior knowledge of DNNs is always unknown. Nevertheless, recent black-box adversarial attacks may struggle to balance their attack ability and visual quality of the generated adversarial examples (AEs) in tackling high-resolution images. In this paper, we propose an attention-guided black-box adversarial attack based on the large-scale multiobjective evolutionary optimization, termed as LMOA. By considering the spatial semantic information of images, we firstly…
Test-Time Backdoor Detection for Object Detection Models
Huazhong University of Science and Technology（华中科技大学）；Harbin Institute of Technology（哈尔滨工业大学）；University of Technology Sydney（悉尼科技大学）；Griffith University（格里菲斯大学）
UFID: A Unified Framework for Input-level Backdoor Detection on Diffusion Models
University of Virginia（弗吉尼亚大学）；Biocomplexity Institute, University of Virginia（弗吉尼亚大学生物复杂性研究所）；School of Data Science, University of Virginia（弗吉尼亚大学数据科学学院）
New Jersey Institute of Technology（新泽西理工学院）；Jordan University of Science & Technology（约旦科技大学）；Qatar Computing Research Institute（卡塔尔计算研究所）；HBKU（哈马德·本·哈利法大学）
Trojan backdoor is a poisoning attack against Neural Network (NN) classifiers in which adversaries try to exploit the (highly desirable) model reuse property to implant Trojans into model parameters for backdoor breaches through a poisoned training process. Most of the proposed defenses against Trojan attacks assume a white-box setup, in which the defender either has access to the inner state of NN or is able to run back-propagation through it. In this work, we propose a more practical black-box defense, dubbed TrojDef, which can only run forward-pass of the NN. TrojDef tries to identify and f…
Black-box Dataset Ownership Verification via Backdoor Watermarking
Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；School of Information Science and Technology, Southwest Jiaotong University（西南交通大学信息科学与技术学院）；Research Center of Artificial Intelligence, Peng Cheng Laboratory（鹏城实验室人工智能研究中心）；Ant Group（蚂蚁集团）
Johns Hopkins University（约翰斯·霍普金斯大学）；University of Maryland（马里兰大学）；New York University（纽约大学）
As the curation of data for machine learning becomes increasingly automated, dataset tampering is a mounting threat. Backdoor attackers tamper with training data to embed a vulnerability in models that are trained on that data. This vulnerability is then activated at inference time by placing a "trigger" into the model's input. Typical backdoor attacks insert the trigger directly into the training data, although the presence of such an attack may be visible upon inspection. In contrast, the Hidden Trigger Backdoor Attack achieves poisoning without placing a trigger into the training data at al…
One major goal of the AI security community is to securely and reliably produce and deploy deep learning models for real-world applications. To this end, data poisoning based backdoor attacks on deep neural networks (DNNs) in the production stage (or training stage) and corresponding defenses are extensively explored in recent years. Ironically, backdoor attacks in the deployment stage, which can often happen in unprofessional users' devices and are thus arguably far more threatening in real-world scenarios, draw much less attention of the community. We attribute this imbalance of vigilance to…
RPP: A Certified Poisoned-Sample Detection Framework for Backdoor Attacks under Dataset Imbalance
Old Dominion University（旧 Dominion 大学）；University of Texas at El Paso（德克萨斯大学埃尔帕索分校）；University of Central Florida（中央佛罗里达大学）；Iowa State University（爱荷华州立大学）；University of Arizona（亚利桑那大学）
NanJing University of Science and Technology（南京理工大学）；University of Adelaide（阿德莱德大学）；Data61, CSIRO（澳大利亚联邦科学与工业研究组织Data61）；Sungkyunkwan University（成均馆大学）
Deep learning models have been shown to be vulnerable to recent backdoor attacks. A backdoored model behaves normally for inputs containing no attacker-secretly-chosen trigger and maliciously for inputs with the trigger. To date, backdoor attacks and countermeasures mainly focus on image classification tasks. And most of them are implemented in the digital world with digital triggers. Besides the classification tasks, object detection systems are also considered as one of the basic foundations of computer vision tasks. However, there is no investigation and understanding of the backdoor vulner…
University of Bucharest（布加勒斯特大学）；University Politehnica of Bucharest（布加勒斯特理工大学）
We study the task of replicating the functionality of black-box neural models, for which we only know the output class probabilities provided for a set of input images. We assume back-propagation through the black-box model is not possible and its training images are not available, e.g. the model could be exposed only through an API. In this context, we present a teacher-student framework that can distill the black-box (teacher) model into a student model with minimal accuracy loss. To generate useful data samples for training the student, our framework (i) learns to generate images on a proxy…
ShrinkBox: Backdoor Attack on Object Detection to Disrupt Collision Avoidance in Machine Learning-based Advanced Driver Assistance Systems
eBRAIN Lab, New York University Abu Dhabi (NYUAD), UAE（eBRAIN实验室，纽约大学阿布扎赫尔分校（NYUAD），阿联酋）；AI and Digital Science Research Center, Technology Innovation Institute (TII), Abu Dhabi, UAE（人工智能与数字科学研究中心，技术创新研究所（TII），阿布扎赫尔，阿联酋）
Deep neural networks (DNNs) are sensitive to adversarial data in a variety of scenarios, including the black-box scenario, where the attacker is only allowed to query the trained model and receive an output. Existing black-box methods for creating adversarial instances are costly, often using gradient estimation or training a replacement network. This paper introduces \textbf{Qu}ery-Efficient \textbf{E}volutiona\textbf{ry} \textbf{Attack}, \textit{QuEry Attack}, an untargeted, score-based, black-box attack. QuEry Attack is based on a novel objective function that can be used in gradient-free o…
The University of Texas at Dallas（德克萨斯大学达拉斯分校）；The Pennsylvania State University（宾夕法尼亚州立大学）
The black-box nature of deep neural networks (DNNs) facilitates attackers to manipulate the behavior of DNN through data poisoning. Being able to detect and mitigate poisoning attacks, typically categorized into backdoor and adversarial poisoning (AP), is critical in enabling safe adoption of DNNs in many application domains. Although recent works demonstrate encouraging results on detection of certain backdoor attacks, they exhibit inherent limitations which may significantly constrain the applicability. Indeed, no technique exists for detecting AP attacks, which represents a harder challenge…
