Search

Adversarial machine learning (AML) is a recent research field that investigates potential security issues related to the use of machine learning (ML) algorithms in modern artificial intelligence (AI)-based systems, along with defensive techniques to protect ML algorithms against such threats. The main threats against ML encompass a set of techniques that aim to mislead ML models through adversarial input perturbations. Unlike ML-enabled crimes, in which ML is used for malicious and offensive purposes, and ML-enabled security mechanisms, in which ML is used for securing existing systems, AML techniques exploit and specifically address the security vulnerabilities of ML algorithms. [ABSTRACT FROM AUTHOR]

Malware still constitutes a major threat in the cybersecurity landscape, also due to the widespread use of infection vectors such as documents. These infection vectors hide embedded malicious code to the victim users, facilitating the use of social engineering techniques to infect their machines. Research showed that machinelearning algorithms provide effective detection mechanisms against such threats, but the existence of an arms race in adversarial settings has recently challenged such systems. In this work, we focus on malware embedded in PDF files as a representative case of such an arms race. We start by providing a comprehensive taxonomy of the different approaches used to generate PDFmalware and of the corresponding learning-based detection systems. We then categorize threats specifically targeted against learning-based PDF malware detectors using a well-established framework in the field of adversarial machine learning. This framework allows us to categorize known vulnerabilities of learning-based PDF malware detectors and to identify novel attacks that may threaten such systems, along with the potential defense mechanisms that can mitigate the impact of such threats. We conclude the article by discussing how such findings highlight promising research directions towards tackling the more general challenge of designing robust malware detectors in adversarial settings. [ABSTRACT FROM AUTHOR]

Learning-based pattern classifiers, including deep networks, have shown impressive performance in several application domains, ranging from computer vision to cybersecurity. However, it has also been shown that adversarial input perturbations carefully crafted either at training or at test time can easily subvert their predictions. The vulnerability of machine learning to such wild patterns (also referred to as adversarial examples), along with the design of suitable countermeasures, have been investigated in the research field of adversarial machine learning. In this work, we provide a thorough overview of the evolution of this research area over the last ten years and beyond, starting from pioneering, earlier work on the security of non-deep learning algorithms up to more recent work aimed to understand the security properties of deep learning algorithms, in the context of computer vision and cybersecurity tasks. We report interesting connections between these apparently-different lines of work, highlighting common misconceptions related to the security evaluation of machine-learning algorithms. We review the main threat models and attacks defined to this end, and discuss the main limitations of current work, along with the corresponding future challenges towards the design of more secure learning algorithms. [ABSTRACT FROM AUTHOR]

Prior work has shown that multibiometric systems are vulnerable to presentation attacks, assuming that their matching score distribution is identical to that of genuine users, without fabricating any fake trait. We have recently shown that this assumption is not representative of current fingerprint and face presentation attacks, leading one to overestimate the vulnerability of multibiometric systems, and to design less effective fusion rules. In this paper, we overcome these limitations by proposing a statistical meta-model of face and fingerprint presentation attacks that characterizes a wider family of fake score distributions, including distributions of known and, potentially, unknown attacks. This allows us to perform a thorough security evaluation of multibiometric systems against presentation attacks, quantifying how their vulnerability may vary also under attacks that are different from those considered during design, through an uncertainty analysis. We empirically show that our approach can reliably predict the performance of multibiometric systems even under never-before-seen face and fingerprint presentation attacks, and that the secure fusion rules designed using our approach can exhibit an improved trade-off between the performance in the absence and in the presence of attack. We finally argue that our method can be extended to other biometrics besides faces and fingerprints. [ABSTRACT FROM AUTHOR]

Machine learning algorithms are increasingly being applied in security-related tasks such as spam and malware detection, although their security properties against deliberate attacks have not yet been widely understood. Intelligent and adaptive attackers may indeed exploit specific vulnerabilities exposed by machine learning techniques to violate system security. Being robust to adversarial data manipulation is thus an important, additional requirement for machine learning algorithms to successfully operate in adversarial settings. In this work, we evaluate the security of Support Vector Machines (SVMs) to well-crafted, adversarial label noise attacks. In particular, we consider an attacker that aims to maximize the SVM׳s classification error by flipping a number of labels in the training data. We formalize a corresponding optimal attack strategy, and solve it by means of heuristic approaches to keep the computational complexity tractable. We report an extensive experimental analysis on the effectiveness of the considered attacks against linear and non-linear SVMs, both on synthetic and real-world datasets. We finally argue that our approach can also provide useful insights for developing more secure SVM learning algorithms, and also novel techniques in a number of related research areas, such as semi-supervised and active learning. [ABSTRACT FROM AUTHOR]

We analyze the problem of designing pattern recognition systems in adversarial settings, under an engineering viewpoint, motivated by their increasing exploitation in security-sensitive applications like spam and malware detection, despite their vulnerability to potential attacks has not yet been deeply understood. We first review previous work and report examples of how a complex system may be evaded either by leveraging on trivial vulnerabilities of its untrained components, e.g. parsing errors in the pre-processing steps, or by exploiting more subtle vulnerabilities of learning algorithms. We then discuss the need of exploiting both reactive and proactive security paradigms complementarily to improve the security by design. Our ultimate goal is to provide some useful guidelines for improving the security of pattern recognition in adversarial settings, and to suggest related open issues to foster research in this area. [ABSTRACT FROM AUTHOR]

Pattern classification systems are commonly used in adversarial applications, like biometric authentication, network intrusion detection, and spam filtering, in which data can be purposely manipulated by humans to undermine their operation. As this adversarial scenario is not taken into account by classical design methods, pattern classification systems may exhibit vulnerabilities, whose exploitation may severely affect their performance, and consequently limit their practical utility. Extending pattern classification theory and design methods to adversarial settings is thus a novel and very relevant research direction, which has not yet been pursued in a systematic way. In this paper, we address one of the main open issues: evaluating at design phase the security of pattern classifiers, namely, the performance degradation under potential attacks they may incur during operation. We propose a framework for empirical evaluation of classifier security that formalizes and generalizes the main ideas proposed in the literature, and give examples of its use in three real applications. Reported results show that security evaluation can provide a more complete understanding of the classifier’s behavior in adversarial environments, and lead to better design choices. [ABSTRACT FROM AUTHOR]

Abstract: In their arms race against developers of spam filters, spammers have recently introduced the image spam trick to make the analysis of emails’ body text ineffective. It consists in embedding the spam message into an attached image, which is often randomly modified to evade signature-based detection, and obfuscated to prevent text recognition by OCR tools. Detecting image spam turns out to be an interesting instance of the problem of content-based filtering of multimedia data in adversarial environments, which is gaining increasing relevance in several applications and media. In this paper we give a comprehensive survey and categorisation of computer vision and pattern recognition techniques proposed so far against image spam, and make an experimental analysis and comparison of some of them on real, publicly available data sets. [Copyright &y& Elsevier]

The increasing digitization and datification of all aspects of people's daily life, and the consequent growth in the use of personal data, are increasingly challenging the current development and adoption of Machine Learning (ML). First, the sheer complexity and amount of data available in these applications strongly demands for ML algorithms that can be trained directly on complex structures, which can be naturally described by graphs. In fact, graphs inherently capture information about entities, their attributes, and relationships between them. Directly applying ML to graphs relieves domain experts and data scientists from the challenging and time-consuming problem of designing a suitable vector-based data representation used by classical ML techniques. Second, ML algorithms should not only be designed to achieve high technical and functional standards; as the automated decisions provided by these algorithms can have a relevant impact on people's lives, their behavior has to be aligned with the values and principles of individuals and society. This demands for designing automated algorithms that we, as humans, can trust, fulfilling the requirements of fairness, robustness, privacy, and explainability. Third, designing effective ML algorithms requires skills and expertise developed at different levels. This substantially hinders the democratization and widespread availability of such technology for society at large, which in turn demands for improving the level of automatization and systematization of their design process, while also providing guarantees on their performance. For this reason, this paper provides an overview of the current works focused towards learning trustworthily, automatically, and with guarantees on graphs. [ABSTRACT FROM AUTHOR]

Adversarial attacks on machine learning-based classifiers, along with defense mechanisms, have been widely studied in the context of single-label classification problems. In this paper, we shift the attention to multi-label classification, where the availability of domain knowledge on the relationships among the considered classes may offer a natural way to spot incoherent predictions, i.e., predictions associated to adversarial examples lying outside of the training data distribution. We explore this intuition in a framework in which first-order logic knowledge is converted into constraints and injected into a semi-supervised learning problem. Within this setting, the constrained classifier learns to fulfill the domain knowledge over the marginal distribution, and can naturally reject samples with incoherent predictions. Even though our method does not exploit any knowledge of attacks during training, our experimental analysis surprisingly unveils that domain-knowledge constraints can help detect adversarial examples effectively, especially if such constraints are not known to the attacker. We show how to implement an adaptive attack exploiting knowledge of the constraints and, in a specifically-designed setting, we provide experimental comparisons with popular state-of-the-art attacks. We believe that our approach may provide a significant step towards designing more robust multi-label classifiers. [ABSTRACT FROM AUTHOR]

Adversarial attacks on machine learning-based classifiers, along with defense mechanisms, have been widely studied in the context of single-label classification problems. In this paper, we shift the attention to multi-label classification, where the availability of domain knowledge on the relationships among the considered classes may offer a natural way to spot incoherent predictions, i.e., predictions associated to adversarial examples lying outside of the training data distribution. We explore this intuition in a framework in which first-order logic knowledge is converted into constraints and injected into a semi-supervised learning problem. Within this setting, the constrained classifier learns to fulfill the domain knowledge over the marginal distribution, and can naturally reject samples with incoherent predictions. Even though our method does not exploit any knowledge of attacks during training, our experimental analysis surprisingly unveils that domain-knowledge constraints can help detect adversarial examples effectively, especially if such constraints are not known to the attacker. We show how to implement an adaptive attack exploiting knowledge of the constraints and, in a specifically-designed setting, we provide experimental comparisons with popular state-of-the-art attacks. We believe that our approach may provide a significant step towards designing more robust multi-label classifiers. [ABSTRACT FROM AUTHOR]

Deep neural networks are vulnerable to adversarial examples, i.e., carefully-crafted inputs that mislead classification at test time. Recent defenses have been shown to improve adversarial robustness by detecting anomalous deviations from legitimate training samples at different layer representations - a behavior normally exhibited by adversarial attacks. Despite technical differences, all aforementioned methods share a common backbone structure that we formalize and highlight in this contribution, as it can help in identifying promising research directions and drawbacks of existing methods. The first main contribution of this work is the review of these detection methods in the form of a unifying framework designed to accommodate both existing defenses and newer ones to come. In terms of drawbacks, the overmentioned defenses require comparing input samples against an oversized number of reference prototypes, possibly at different representation layers, dramatically worsening the test-time efficiency. Besides, such defenses are typically based on ensembling classifiers with heuristic methods, rather than optimizing the whole architecture in an end-to-end manner to better perform detection. As a second main contribution of this work, we introduce FADER, a novel technique for speeding up detection-based methods. FADER overcome the issues above by employing RBF networks as detectors: by fixing the number of required prototypes, the runtime complexity of adversarial examples detectors can be controlled. Our experiments outline up to 73 × prototypes reduction compared to analyzed detectors for MNIST dataset, up to 50 × for CIFAR10 dataset, and up to 82 × on ImageNet10 dataset respectively, without sacrificing classification accuracy on both clean and adversarial data. [ABSTRACT FROM AUTHOR]

• We provide a dataset to benchmark machine-learning models against adversarial patches. • This dataset enables approximating models' robustness, avoiding a computationally and time-demanding evaluation. • We tested the effectiveness of the generated adversarial patches against 127 models. • We showed that the generated adversarial patches are effective when printed and applied to real-world objects. • We open-source the code to evaluate models robustness using our dataset. Adversarial patches are optimized contiguous pixel blocks in an input image that cause a machine-learning model to misclassify it. However, their optimization is computationally demanding, and requires careful hyperparameter tuning, potentially leading to suboptimal robustness evaluations. To overcome these issues, we propose ImageNet-Patch, a dataset to benchmark machine-learning models against adversarial patches. The dataset is built by first optimizing a set of adversarial patches against an ensemble of models, using a state-of-the-art attack that creates transferable patches. The corresponding patches are then randomly rotated and translated, and finally applied to the ImageNet data. We use ImageNet-Patch to benchmark the robustness of 127 models against patch attacks, and also validate the effectiveness of the given patches in the physical domain (i.e., by printing and applying them to real-world objects). We conclude by discussing how our dataset could be used as a benchmark for robustness, and how our methodology can be generalized to other domains. We open source our dataset and evaluation code at https://github.com/pralab/ImageNet-Patch. [ABSTRACT FROM AUTHOR]