Pixel Logo Attack: Embedding Attacks as Logo-Like Pixels

Adversarial example generation is a technique that involves perturbing inputs with imperceptible noise to induce misclassifications in neural networks, serving as a means to assess the robustness of such models. Among the adversarial attack algorithms, momentum iterative fast gradient sign Method (MI-FGSM) and its variants constitute a class of highly effective offensive strategies, achieving near-perfect attack success rates in white-box settings. However, these methods’ use of sign activation functions severely compromises gradient information, which leads to low success rates in black-box attacks and results in large adversarial perturbations. In this paper, we introduce a novel adversarial attack algorithm, NA-FGTM. Our method employs the Tanh activation function instead of the sign which can accurately preserve gradient information. In addition, it utilizes the Adam optimization algorithm as well as the Nesterov acceleration, which is able to stabilize gradient update directions and expedite gradient convergence. Above all, the transferability of adversarial examples can be enhanced. Through integration with data augmentation techniques such as DIM, TIM, and SIM, NA-FGTM can further improve the efficacy of black-box attacks. Extensive experiments on the ImageNet dataset demonstrate that our method outperforms the state-of-the-art approaches in terms of black-box attack success rate and generates adversarial examples with smaller perturbations.

Recent studies have shown that deep learning models such as image classification and object detection models are vulnerable to attacks from adversarial examples. These malicious examples may trigger potential security threats and cause damage in security-related fields. In this paper, in order to find model robustness weakness, we obtain the adversarial robustness metric against several adversarial attacks, including gradient-based attacks, optimization-based attacks, etc. The metric quantizes the adversarial robustness of deep learning models and can be used as a guide for extra model training. Meanwhile, we design a testing platform to achieve model adversarial robustness evaluation. Utilizing the front and back end separation strategy, the testing platform consist of 2 modules: one is multiple adversarial example attacks executor and the other is HCI module. In addition, the tool allows parallel comparison among different adversarial attack algorithms under given conditions and measures the quality of generated adversarial examples utilizing FID(Fréchet Inception Distance) metrics. Besides, we developed ERFGSM(edge-RFGSM), a new gradient-based attack using edge information through canny operator.

Deep neural networks (DNNs) are vulnerable to adversarial examples (AEs) that mislead the model while appearing benign to human observers. A critical concern is the transferability of AEs, which enables black-box attacks without direct access to the target model. However, many previous attacks have failed to explain the intrinsic mechanism of adversarial transferability, lacking a unified and representative metric for transferability as well. In this paper, we rethink the property of transferable AEs and develop a novel metric to measure transferability from the perspective of generalization. Building on insights from this metric, we analyze the generalization of AEs across models with different architectures and prove that we can find a local perturbation to mitigate the gap between surrogate and target models. We further establish the inner connections between model smoothness and flat local maxima, both of which contribute to the transferability of AEs. Further, we propose a new adversarial attack algorithm, Adversarial Weight Tuning (AWT), which adaptively adjusts the parameters of the surrogate model using generated AEs to optimize the flat local maxima and model smoothness simultaneously, without the need for extra data. AWT is a data-free tuning method that combines gradient-based and model-related attack methods to enhance the transferability of AEs. Extensive experiments on a variety of models with different architectures on ImageNet demonstrate that AWT yields superior performance over other attacks, with an average increase of nearly 5% and 10% attack success rates on CNN-based and Transformer-based models, respectively, compared to state-of-the-art attacks.

Although deep neural networks (DNNs) have achieved high performance across various applications, they are often deceived by adversarial examples generated by adding small perturbations. To combat adversarial attacks, many detection methods have been proposed, including feature squeezing and trapdoor. However, these methods rely on the output of DNNs or involve training a separate network to detect adversarial examples, which leads to high computational costs and low efficiency. In this study, we propose a simple and effective approach called the entropy-based detector (EBD) to protect DNNs from various adversarial attacks. EBD detects adversarial examples by comparing the difference in entropy between the input sample before and after bit depth reduction. We show that EBD can detect over 98% of the adversarial examples generated by attacks using fast-gradient sign method, basic iterative method, momentum iterative method, DeepFool and CW attacks when the false positive rate is 2.5% for CIFAR-10 and ImageNet datasets.

Adversarial examples are one of the largest vulnerability of deep neural networks. An attacker can deceive the classifiers easily with the malicious inputs (called adversarial examples), which perturbations are slightly added to benign inputs. Various attack methods have been studied in both white-box and black-box settings, and some methods achieve high attack success rates even in the black-box settings; that is, the attacker is restricted to only query accesses to the target network. In this paper, we propose a simple hyperparameter-free score-based black-box \(\ell _\infty \)-adversarial attack using local uniform noises and a random search. Specifically, we construct adversarial perturbations by combining local uniform noises such as vertical-wise and horizontal-wise, and incorporate this idea into the random search method to update the perturbation sequentially. We evaluate our method in terms of attack success rates and query efficiency using models that classify common datasets CIFAR-10 and ImageNet. We show that our method achieves higher attack success rates and query efficiency than previous attack methods, especially in low-query budgets on both untargeted and targeted attack settings. We also examine attacks to adversarially trained models and discuss the effect of local uniform noises on these models. Furthermore, we show that our method achieves relatively high attack success rates and query efficiency on average against input-transformation-based defense methods, and is virtually unaffected by these defense methods.KeywordsBlack-box adversarial attacksAI security

In recent years, deep learning has been increasingly used to the detection of pests and diseases. Unfortunately, deep neural networks are particularly vulnerable when attacked by adversarial examples. Hence it is vital to explore the creation of intensely aggressive adversarial examples to increase neural network robustness. This paper proposes a wavelet transform and histogram equalization-based adversarial attack algorithm: WT-MI-FGSM. In order to verify the performance of the WT-MI-FGSM, we propose a plant pests and diseases identification method based on the coordinate attention mechanism and CondenseNetV2: CL-CondenseNetV2. The accuracy of CL- CondenseNetV2 on the PlantVillage dataset is 99.45%, which indicates that the improved CondenseNetV2 model has a more significant classification performance. In adversarial sample experiments using WT-MI-FGSM and CL-CondenseNetV2, experimental results show that when CL-CondenseNetV2 is attacked by the adversarial algorithm WT-MI-FGSM, the error rate reaches 89.8%, with a higher attack success rate than existing adversarial attack algorithms. In addition, the accuracy of CL-CondenseNetV2 is improved to 99.71% by adding the adversarial samples generated by WT-MI-FGSM to the training set and performing adversarial training. The experiments demonstrate that the adversarial examples caused by WT-MI-FGSM can improve the model's performance.

Identifying objects in surveillance and reconnaissance systems with the human eye can be challenging, underscoring the growing importance of employing deep learning models for the recognition of enemy weapon systems. These systems, leveraging deep neural networks known for their strong performance in image recognition and classification, are currently under extensive research. However, it is crucial to acknowledge that surveillance and reconnaissance systems utilizing deep neural networks are susceptible to vulnerabilities posed by adversarial examples. While prior adversarial example research has mainly utilized publicly available internet data, there has been a significant absence of studies concerning adversarial attacks on data and models specific to real military scenarios. In this paper, we introduce an adversarial example designed for a binary classifier tasked with recognizing helicopters. Our approach generates an adversarial example that is misclassified by the model, despite appearing unproblematic to the human eye. To conduct our experiments, we gathered real attack and transport helicopters and employed TensorFlow as the machine learning library of choice. Our experimental findings demonstrate that the average attack success rate of the proposed method is 81.9%. Additionally, when epsilon is 0.4, the attack success rate is 90.1%. Before epsilon reaches 0.4, the attack success rate increases rapidly, and then we can see that epsilon increases little by little thereafter.

Some recent articles have revealed that synthetic aperture radar automatic target recognition (SAR-ATR) models based on deep learning are vulnerable to the attacks of adversarial examples and cause security problems. The adversarial attack can make a deep convolutional neural network (CNN)-based SAR-ATR system output the intended wrong label predictions by adding small adversarial perturbations to the SAR images. The existing optimization-based adversarial attack methods generate adversarial examples by minimizing the mean-squared reconstruction error, causing smooth target edge and blurry weak scattering centers in SAR images. In this paper, we build a UNet-generative adversarial network (GAN) to refine the generation of the SAR-ATR models’ adversarial examples. The UNet learns the separable features of the targets and generates the adversarial examples of SAR images. The GAN makes the generated adversarial examples approximate to real SAR images (with sharp target edge and explicit weak scattering centers) and improves the generation efficiency. We carry out abundant experiments using the proposed adversarial attack algorithm to fool the SAR-ATR models based on several advanced CNNs, which are trained on the measured SAR images of the ground vehicle targets. The quantitative and qualitative results demonstrate the high-quality adversarial example generation and excellent attack effectiveness and efficiency improvement.

Explanation-Guided Adversarial Example Attacks

Adversarial attacks expose the latent vulnerabilities within artificial intelligence systems, necessitating a reassessment and enhancement of model robustness to ensure the reliability and security of deep learning models against malicious attacks. We propose a fast method designed to efficiently find sample points close to the decision boundary. By computing the gradient information of each class in the input samples and comparing these gradient differences with the true class, we can identify the target class most sensitive to the decision boundary, thus generating adversarial examples. This technique is referred to as the “You Only Attack Once” (YOAO) algorithm. Compared to the DeepFool algorithm, this method requires only a single iteration to achieve effective attack results. The experimental results demonstrate that the proposed algorithm outperforms the original approach in various scenarios, especially in resource-constrained environments. Under a single iteration, it achieves a 70.6% higher success rate of the attacks compared to the DeepFool algorithm. Our proposed method shows promise for widespread application in both offensive and defensive strategies for diverse deep learning models. We investigated the relationship between classifier accuracy and adversarial attack success rate, comparing the algorithm with others. Our experiments validated that the proposed algorithm exhibits higher attack success rates and efficiency. Furthermore, we performed data visualization on the ImageNet dataset, demonstrating that the proposed algorithm focuses more on attacking important features. Finally, we discussed the existing issues with the algorithm and outlined future research directions. Our code will be made public upon acceptance of the paper.

The adversarial attack is a cutting-edge technique used to study the vulnerability of deep neural networks (DNNs). However, most existing studies focus on the additive perturbation-based attack, which cannot represent real-world corruption and limit their applications. In particular, haze is a common natural phenomenon that significantly corrupts an image, which inevitably poses a potential threat to deep models. In this work, for the first attempt, we study the effects of haze on DNNs from the perspective of adversarial attacks and propose two adversarial haze attack methods. We first propose the optimization-based adversarial haze attack (OAdvHaze) that optimizes the parameters of the atmospheric scattering model with the guidance of a DNN to synthesize a hazy image, which leads to a high attack success rate. To achieve a more efficient attack, we further propose a predictive adversarial haze attack (PAdvHaze) employing a DNN to predict the hazing parameters through a one-step way. To validate the effectiveness of both methods, we conducted extensive experiments on two publicly available datasets, i.e., ILSVRC 2012 and NIPS 2017. OAdvHaze and PAdvHaze achieve comparable attack success rates and transferability to state-of-the-art attacks. This work would contribute to the evaluation and enhancement of the robustness of DNNs against haze perturbation that may happen in the real world.

Adversarial examples are perturbed inputs that are designed (from a deep learning network's (DLN) parameter gradients) to mislead the DLN during test time. Intuitively, constraining the dimensionality of inputs or parameters of a network reduces the 'space' in which adversarial examples exist. Guided by this intuition, we demonstrate that discretization greatly improves the robustness of DLNs against adversarial attacks. Specifically, discretizing the input space (or allowed pixel levels from 256 values or 8-bit to 4 values or 2-bit) extensively improves the adversarial robustness of DLNs for a substantial range of perturbations for minimal loss in test accuracy. Furthermore, we find that Binary Neural Networks (BNNs) and related variants are intrinsically more robust than their full precision counterparts in adversarial scenarios. Combining input discretization with BNNs furthers the robustness even waiving the need for adversarial training for certain magnitude of perturbation values. We evaluate the effect of discretization on MNIST, CIFAR10, CIFAR100 and Imagenet datasets. Across all datasets, we observe maximal adversarial resistance with 2-bit input discretization that incurs an adversarial accuracy loss of just ~1-2% as compared to clean test accuracy.

Synthetic aperture radar (SAR) automatic target recognition (ATR) models based on deep neural networks (DNNs) are susceptible to adversarial attacks. In this study, we proposed an SAR black-box local adversarial attack algorithm named attention heat map- based black-box local adversarial attack (AH-BLAA). First, we designed an attention heat map extraction module combined with the layer-wise relevance propagation (LRP) algorithm to obtain the high concerning areas of the SAR-ATR models. Then, to gener- ate SAR adversarial attack examples, we designed a perturbation generator module, introducing the structural dissimilarity (DSSIM) metric in the loss function to limit image distortion and the dif- ferential evolution (DE) algorithm to search for optimal perturba- tions. Experimental results on the MSTAR and FUSAR-Ship datasets showed that compared with existing adversarial attack algorithms, the attack success rate of the AH-BLAA algorithm increased by 0.63% to 33.59% and 1.05% to 17.65%, respectively. Moreover, the low- est perturbation ratios reached 0.23% and 0.13%, respectively.

FE-DaST: Fast and effective data-free substitute training for black-box adversarial attacks

DiffProtect: Generative adversarial examples using diffusion models for facial privacy protection