Biblio

Differential privacy has emerged as a formal framework for protecting sensitive information in control systems. One key feature is that it is immune to post-processing, which means that arbitrary post-hoc computations can be performed on privatized data without weakening differential privacy. It is therefore common to filter private data streams. To characterize this setup, in this paper we present error and entropy bounds for Kalman filtering differentially private state trajectories. We consider systems in which an output trajectory is privatized in order to protect the state trajectory that produced it. We provide bounds on a priori and a posteriori error and differential entropy of a Kalman filter which is processing the privatized output trajectories. Using the error bounds we develop, we then provide guidelines to calibrate privacy levels in order to keep filter error within pre-specified bounds. Simulation results are presented to demonstrate these developments.

Conventional adversarial defenses reduce classification accuracy whether or not a model is under attacks. Moreover, most of image processing based defenses are defeated due to the problem of obfuscated gradients. In this paper, we propose a new adversarial defense which is a defensive transform for both training and test images inspired by perceptual image encryption methods. The proposed method utilizes a block-wise pixel shuffling method with a secret key. The experiments are carried out on both adaptive and non-adaptive maximum-norm bounded white-box attacks while considering obfuscated gradients. The results show that the proposed defense achieves high accuracy (91.55%) on clean images and (89.66%) on adversarial examples with noise distance of 8/255 on CFAR-10 dataset. Thus, the proposed defense outperforms state-of-the-art adversarial defenses including latent adversarial training, adversarial training and thermometer encoding.

Neural networks with deep architectures have been used to construct state-of-the-art classifiers that can match human level accuracy in areas such as image classification. However, many of these classifiers can be fooled by examples slightly modified from their original forms. In this work, we propose a novel approach for generating adversarial examples that makes use of only attribution information of the features and perturbs only features that are highly influential to the output of the classifier. We call this approach Attribution Based Adversarial Generation (ABAG). To demonstrate the effectiveness of this approach, three somewhat arbitrary algorithms are proposed and examined. In the first algorithm all non-zero attributions are utilized and associated features perturbed; in the second algorithm only the top-n most positive and top-n most negative attributions are used and corresponding features perturbed; and in the third algorithm the level of perturbation is increased in an iterative manner until an adversarial example is discovered. All of the three algorithms are implemented and experiments are performed on the well-known MNIST dataset. Experiment results show that adversarial examples can be generated very efficiently, and thus prove the validity and efficacy of ABAG - utilizing attributions for the generation of adversarial examples. Furthermore, as shown by examples, ABAG can be adapted to provides a systematic searching approach to generate adversarial examples by perturbing a minimum amount of features.

We study the attribution problem for deep networks applied to perception tasks. For vision tasks, attribution techniques attribute the prediction of a network to the pixels of the input image. We propose a new technique called Blur Integrated Gradients (Blur IG). This technique has several advantages over other methods. First, it can tell at what scale a network recognizes an object. It produces scores in the scale/frequency dimension, that we find captures interesting phenomena. Second, it satisfies the scale-space axioms, which imply that it employs perturbations that are free of artifact. We therefore produce explanations that are cleaner and consistent with the operation of deep networks. Third, it eliminates the need for baseline parameter for Integrated Gradients for perception tasks. This is desirable because the choice of baseline has a significant effect on the explanations. We compare the proposed technique against previous techniques and demonstrate application on three tasks: ImageNet object recognition, Diabetic Retinopathy prediction, and AudioSet audio event identification. Code and examples are at https://github.com/PAIR-code/saliency.

Spiking Neural Networks (SNNs) claim to present many advantages in terms of biological plausibility and energy efficiency compared to standard Deep Neural Networks (DNNs). Recent works have shown that DNNs are vulnerable to adversarial attacks, i.e., small perturbations added to the input data can lead to targeted or random misclassifications. In this paper, we aim at investigating the key research question: "Are SNNs secure?" Towards this, we perform a comparative study of the security vulnerabilities in SNNs and DNNs w.r.t. the adversarial noise. Afterwards, we propose a novel black-box attack methodology, i.e., without the knowledge of the internal structure of the SNN, which employs a greedy heuristic to automatically generate imperceptible and robust adversarial examples (i.e., attack images) for the given SNN. We perform an in-depth evaluation for a Spiking Deep Belief Network (SDBN) and a DNN having the same number of layers and neurons (to obtain a fair comparison), in order to study the efficiency of our methodology and to understand the differences between SNNs and DNNs w.r.t. the adversarial examples. Our work opens new avenues of research towards the robustness of the SNNs, considering their similarities to the human brain's functionality.

Deep Learning (DL), in spite of its huge success in many new fields, is extremely vulnerable to adversarial attacks. We demonstrate how an attacker applies physical white-box and black-box adversarial attacks to Channel decoding systems based on DL. We show that these attacks can affect the systems and decrease performance. We uncover that these attacks are more effective than conventional jamming attacks. Additionally, we show that classical decoding schemes are more robust than the deep learning channel decoding systems in the presence of both adversarial and jamming attacks.

Adversarial examples are known as carefully perturbed images fooling image classifiers. We propose a geometric framework to generate adversarial examples in one of the most challenging black-box settings where the adversary can only generate a small number of queries, each of them returning the top-1 label of the classifier. Our framework is based on the observation that the decision boundary of deep networks usually has a small mean curvature in the vicinity of data samples. We propose an effective iterative algorithm to generate query-efficient black-box perturbations with small p norms which is confirmed via experimental evaluations on state-of-the-art natural image classifiers. Moreover, for p=2, we theoretically show that our algorithm actually converges to the minimal perturbation when the curvature of the decision boundary is bounded. We also obtain the optimal distribution of the queries over the iterations of the algorithm. Finally, experimental results confirm that our principled black-box attack algorithm performs better than state-of-the-art algorithms as it generates smaller perturbations with a reduced number of queries.

Although deep convolutional neural network (CNN) performs well in many computer vision tasks, its classification mechanism is very vulnerable when it is exposed to the perturbation of adversarial attacks. In this paper, we proposed a new algorithm to generate the substitute model of black-box CNN models by using knowledge distillation. The proposed algorithm distills multiple CNN teacher models to a compact student model as the substitution of other black-box CNN models to be attacked. The black-box adversarial samples can be consequently generated on this substitute model by using various white-box attacking methods. According to our experiments on ResNet18 and DenseNet121, our algorithm boosts the attacking success rate (ASR) by 20% by training the substitute model based on knowledge distillation.

It is now possible to synthesize highly realistic images of people who do not exist. Such content has, for example, been implicated in the creation of fraudulent socialmedia profiles responsible for dis-information campaigns. Significant efforts are, therefore, being deployed to detect synthetically-generated content. One popular forensic approach trains a neural network to distinguish real from synthetic content.We show that such forensic classifiers are vulnerable to a range of attacks that reduce the classifier to near- 0% accuracy. We develop five attack case studies on a state- of-the-art classifier that achieves an area under the ROC curve (AUC) of 0.95 on almost all existing image generators, when only trained on one generator. With full access to the classifier, we can flip the lowest bit of each pixel in an image to reduce the classifier's AUC to 0.0005; perturb 1% of the image area to reduce the classifier's AUC to 0.08; or add a single noise pattern in the synthesizer's latent space to reduce the classifier's AUC to 0.17. We also develop a black-box attack that, with no access to the target classifier, reduces the AUC to 0.22. These attacks reveal significant vulnerabilities of certain image-forensic classifiers.

Deep Neural Networks (DNNs) are susceptible to model stealing attacks, which allows a data-limited adversary with no knowledge of the training dataset to clone the functionality of a target model, just by using black-box query access. Such attacks are typically carried out by querying the target model using inputs that are synthetically generated or sampled from a surrogate dataset to construct a labeled dataset. The adversary can use this labeled dataset to train a clone model, which achieves a classification accuracy comparable to that of the target model. We propose "Adaptive Misinformation" to defend against such model stealing attacks. We identify that all existing model stealing attacks invariably query the target model with Out-Of-Distribution (OOD) inputs. By selectively sending incorrect predictions for OOD queries, our defense substantially degrades the accuracy of the attacker's clone model (by up to 40%), while minimally impacting the accuracy (\textbackslashtextless; 0.5%) for benign users. Compared to existing defenses, our defense has a significantly better security vs accuracy trade-off and incurs minimal computational overhead.

Artificial neural networks in general and deep learning networks in particular established themselves as popular and powerful machine learning algorithms. While the often tremendous sizes of these networks are beneficial when solving complex tasks, the tremendous number of parameters also causes such networks to be vulnerable to malicious behavior such as adversarial perturbations. These perturbations can change a model's classification decision. Moreover, while single-step adversaries can easily be transferred from network to network, the transfer of more powerful multi-step adversaries has - usually - been rather difficult.In this work, we introduce a method for generating strong adversaries that can easily (and frequently) be transferred between different models. This method is then used to generate a large set of adversaries, based on which the effects of selected defense methods are experimentally assessed. At last, we introduce a novel, simple, yet effective approach to enhance the resilience of neural networks against adversaries and benchmark it against established defense methods. In contrast to the already existing methods, our proposed defense approach is much more efficient as it only requires a single additional forward-pass to achieve comparable performance results.

Increasing deep neural networks are applied in solving graph evolved tasks, such as node classification and link prediction. However, the vulnerability of deep models can be revealed using carefully crafted adversarial examples generated by various adversarial attack methods. To explore this security problem, we define the link prediction adversarial attack problem and put forward a novel iterative gradient attack (IGA) strategy using the gradient information in the trained graph autoencoder (GAE) model. Not surprisingly, GAE can be fooled by an adversarial graph with a few links perturbed on the clean one. The results on comprehensive experiments of different real-world graphs indicate that most deep models and even the state-of-the-art link prediction algorithms cannot escape the adversarial attack, such as GAE. We can benefit the attack as an efficient privacy protection tool from the link prediction of unknown violations. On the other hand, the adversarial attack is a robust evaluation metric for current link prediction algorithms of their defensibility.

This work uses adversarial perturbations to enhance deepfake images and fool common deepfake detectors. We created adversarial perturbations using the Fast Gradient Sign Method and the Carlini and Wagner L2 norm attack in both blackbox and whitebox settings. Detectors achieved over 95% accuracy on unperturbed deepfakes, but less than 27% accuracy on perturbed deepfakes. We also explore two improvements to deep-fake detectors: (i) Lipschitz regularization, and (ii) Deep Image Prior (DIP). Lipschitz regularization constrains the gradient of the detector with respect to the input in order to increase robustness to input perturbations. The DIP defense removes perturbations using generative convolutional neural networks in an unsupervised manner. Regularization improved the detection of perturbed deepfakes on average, including a 10% accuracy boost in the blackbox case. The DIP defense achieved 95% accuracy on perturbed deepfakes that fooled the original detector while retaining 98% accuracy in other cases on a 100 image subsample.

In this paper, differential privacy protection method is applied to matrix factorization method that used to solve the recommendation problem. For centralized recommendation scenarios, a collaborative filtering recommendation model based on matrix factorization is established, and a matrix factorization mechanism satisfying ε-differential privacy is proposed. Firstly, the potential characteristic matrix of users and projects is constructed. Secondly, noise is added to the matrix by the method of target disturbance, which satisfies the differential privacy constraint, then the noise matrix factorization model is obtained. The parameters of the model are obtained by the stochastic gradient descent algorithm. Finally, the differential privacy matrix factorization model is used for score prediction. The effectiveness of the algorithm is evaluated on the public datasets including Movielens and Netflix. The experimental results show that compared with the existing typical recommendation methods, the new matrix factorization method with privacy protection can recommend within a certain range of recommendation accuracy loss while protecting the users' privacy information.

Recently, the field of adversarial machine learning has been garnering attention by showing that state-of-the-art deep neural networks are vulnerable to adversarial examples, stemming from small perturbations being added to the input image. Adversarial examples are generated by a malicious adversary by obtaining access to the model parameters, such as gradient information, to alter the input or by attacking a substitute model and transferring those malicious examples over to attack the victim model. Specifically, one of these attack algorithms, Robust Physical Perturbations (RP2), generates adversarial images of stop signs with black and white stickers to achieve high targeted misclassification rates against standard-architecture traffic sign classifiers. In this paper, we propose BlurNet, a defense against the RP2 attack. First, we motivate the defense with a frequency analysis of the first layer feature maps of the network on the LISA dataset, which shows that high frequency noise is introduced into the input image by the RP2 algorithm. To remove the high frequency noise, we introduce a depthwise convolution layer of standard blur kernels after the first layer. We perform a blackbox transfer attack to show that low-pass filtering the feature maps is more beneficial than filtering the input. We then present various regularization schemes to incorporate this lowpass filtering behavior into the training regime of the network and perform white-box attacks. We conclude with an adaptive attack evaluation to show that the success rate of the attack drops from 90% to 20% with total variation regularization, one of the proposed defenses.

Partitional Clustering Algorithm (PCA) on the Hadoop Distributed File System is to perform big data securities using the Perturbation Technique is the main idea of the proposed work. There are numerous clustering methods available that are used to categorize the information from the big data. PCA discovers the cluster based on the initial partition of the data. In this approach, it is possible to develop a security safeguarding of data that is impoverished to allow the calculations and communication. The performances were analyzed on Health Care database under the studies of various parameters like precision, accuracy, and F-score measure. The outcome of the results is to demonstrate that this method is used to decrease the complication in preserving privacy and better accuracy than that of the existing techniques.

The Robot Operating System (ROS) are being deployed for multiple life critical activities such as self-driving cars, drones, and industries. However, the security has been persistently neglected, especially the image flows incoming from camera robots. In this paper, we perform a structured security assessment of robot cameras using ROS. We points out a relevant number of security flaws that can be used to take over the flows incoming from the robot cameras. Furthermore, we propose an intrusion detection system to detect abnormal flows. Our defense approach is based on images comparisons and unsupervised anomaly detection method. We experiment our approach on robot cameras embedded on a self-driving car.

With the increasing performance of neural networks on many security-critical tasks, the security concerns of machine learning have become increasingly prominent. Recent studies have shown that neural networks are vulnerable to adversarial examples: carefully crafted inputs with negligible perturbations on legitimate samples could mislead a neural network to produce adversary-selected outputs while humans can still correctly classify them. Therefore, we need an additional measurement on the trustworthiness of the results of a machine learning model, especially in adversarial settings. In this paper, we analyse the root cause of adversarial examples, and propose a new property, namely fidelity, of machine learning models to describe the gap between what a model learns and the ground truth learned by humans. One of its benefits is detecting adversarial attacks. We formally define fidelity, and propose a novel approach to quantify it. We evaluate the quantification of fidelity in adversarial settings on two neural networks. The study shows that involving the fidelity enables a neural network system to detect adversarial examples with true positive rate 97.7%, and false positive rate 1.67% on a studied neural network.

Intrusion detection systems (IDSs) are an essential cog of the network security suite that can defend the network from malicious intrusions and anomalous traffic. Many machine learning (ML)-based IDSs have been proposed in the literature for the detection of malicious network traffic. However, recent works have shown that ML models are vulnerable to adversarial perturbations through which an adversary can cause IDSs to malfunction by introducing a small impracticable perturbation in the network traffic. In this paper, we propose an adversarial ML attack using generative adversarial networks (GANs) that can successfully evade an ML-based IDS. We also show that GANs can be used to inoculate the IDS and make it more robust to adversarial perturbations.

Internet application providers now have more incentive than ever to collect user data, which greatly increases the risk of user privacy violations due to the emerging of deep neural networks. In this paper, we propose TensorClog-a poisoning attack technique that is designed for privacy protection against deep neural networks. TensorClog has three properties with each of them serving a privacy protection purpose: 1) training on TensorClog poisoned data results in lower inference accuracy, reducing the incentive of abusive data collection; 2) training on TensorClog poisoned data converges to a larger loss, which prevents the neural network from learning the privacy; and 3) TensorClog regularizes the perturbation to remain a high structure similarity, so that the poisoning does not affect the actual content in the data. Applying our TensorClog poisoning technique to CIFAR-10 dataset results in an increase in both converged training loss and test error by 300% and 272%, respectively. It manages to maintain data's human perception with a high SSIM index of 0.9905. More experiments including different limited information attack scenarios and a real-world application transferred from pre-trained ImageNet models are presented to further evaluate TensorClog's effectiveness in more complex situations.

Malicious software, known as malware, has become urgently serious threat for computer security, so automatic mal-ware classification techniques have received increasing attention. In recent years, deep learning (DL) techniques for computer vision have been successfully applied for malware classification by visualizing malware files and then using DL to classify visualized images. Although DL-based classification systems have been proven to be much more accurate than conventional ones, these systems have been shown to be vulnerable to adversarial attacks. However, there has been little research to consider the danger of adversarial attacks to visualized image-based malware classification systems. This paper proposes an adversarial attack method based on the gradient to attack image-based malware classification systems by introducing perturbations on resource section of PE files. The experimental results on the Malimg dataset show that by a small interference, the proposed method can achieve success attack rate when challenging convolutional neural network malware classifiers.

The use of biometric-based authentication systems spread over daily life consumer electronics. Over the years, researchers' interest shifted from hard (such as fingerprints, voice and keystroke dynamics) to soft biometrics (such as age, ethnicity and gender), mainly by using the latter to improve the authentication systems effectiveness. While newer approaches are constantly being proposed by domain experts, in the last years Deep Learning has raised in many computer vision tasks, also becoming the current state-of-art for several biometric approaches. However, since the automatic processing of data rich in sensitive information could expose users to privacy threats associated to their unfair use (i.e. gender or ethnicity), in the last years researchers started to focus on the development of defensive strategies in the view of a more secure and private AI. The aim of this work is to exploit Adversarial Perturbation, namely approaches able to mislead state-of-the-art CNNs by injecting a suitable small perturbation over the input image, to protect subjects against unwanted soft biometrics-based identification by automatic means. In particular, since ethnicity is one of the most critical soft biometrics, as a case of study we will focus on the generation of adversarial stickers that, once printed, can hide subjects ethnicity in a real-world scenario.

The article analyzes the concept of "Resilience" in relation to the development of computing. The strategy for reacting to perturbations in this process can be based either on "harsh Resistance" or "smarter Elasticity." Our "Models" are descriptive in defining the path of evolutionary development as structuring under the perturbations of the natural order and enable the analysis of the relationship among models, structures and factors of evolution. Among those, two features are critical: parallelism and "fuzziness", which to a large extent determine the rate of change of computing development, especially in critical applications. Both reversible and irreversible development processes related to elastic and resistant methods of problem solving are discussed. The sources of perturbations are located in vicinity of the resource boundaries, related to growing problem size with progress combined with the lack of computational "checkability" of resources i.e. data with inadequate models, methodologies and means. As a case study, the problem of hidden faults caused by the growth, the deficit of resources, and the checkability of digital circuits in critical applications is analyzed.

With rapid development of deep integration of computation, control, and communication, Cyber-Physical Systems (CPSs) play an important role in industrial processes. Combined with the technology of fog computing, CPSs can outsource their complicated computation to the fog layer, which in turn, may bring security threats with regard to data privacy. To protect data privacy in a control framework, this paper investigate observer-based secure control problem towards fog computing aided CPSs (FCA-CPSs) by utilizing data perturbation method. Firstly, security inputs are designed to encrypt the transmitted states to realize specific confidentiality level. Then, sufficient conditions are established to ensure the stability of considered FCA-CPSs. Finally, a numerical example is provided to illustrate the effectiveness of the secure estimation scheme.

Alternating direction method of multiplier (ADMM) is a powerful method to solve decentralized convex optimization problems. In distributed settings, each node performs computation with its local data and the local results are exchanged among neighboring nodes in an iterative fashion. During this iterative process the leakage of data privacy arises and can accumulate significantly over many iterations, making it difficult to balance the privacy-utility tradeoff. In this study we propose Recycled ADMM (R-ADMM), where a linear approximation is applied to every even iteration, its solution directly calculated using only results from the previous, odd iteration. It turns out that under such a scheme, half of the updates incur no privacy loss and require much less computation compared to the conventional ADMM. We obtain a sufficient condition for the convergence of R-ADMM and provide the privacy analysis based on objective perturbation.