Carl A. Gunter

Xiaojun Xu, Yue Yu, Hanzhang Wang et al.

Graph Neural Networks (GNNs) have been widely applied to different tasks such as bioinformatics, drug design, and social networks. However, recent studies have shown that GNNs are vulnerable to adversarial attacks which aim to mislead the node or subgraph classification prediction by adding subtle perturbations. Detecting these attacks is challenging due to the small magnitude of perturbation and the discrete nature of graph data. In this paper, we propose a general adversarial edge detection pipeline EDoG without requiring knowledge of the attack strategies based on graph generation. Specifically, we propose a novel graph generation approach combined with link prediction to detect suspicious adversarial edges. To effectively train the graph generative model, we sample several sub-graphs from the given graph data. We show that since the number of adversarial edges is usually low in practice, with low probability the sampled sub-graphs will contain adversarial edges based on the union bound. In addition, considering the strong attacks which perturb a large number of edges, we propose a set of novel features to perform outlier detection as the preprocessing for our detection. Extensive experimental results on three real-world graph datasets including a private transaction rule dataset from a major company and two types of synthetic graphs with controlled properties show that EDoG can achieve above 0.8 AUC against four state-of-the-art unseen attack strategies without requiring any knowledge about the attack type; and around 0.85 with knowledge of the attack type. EDoG significantly outperforms traditional malicious edge detection baselines. We also show that an adaptive attack with full knowledge of our detection pipeline is difficult to bypass it.

Yunhui Long, Boxin Wang, Zhuolin Yang et al.

Recent advances in machine learning have largely benefited from the massive accessible training data. However, large-scale data sharing has raised great privacy concerns. In this work, we propose a novel privacy-preserving data Generative model based on the PATE framework (G-PATE), aiming to train a scalable differentially private data generator that preserves high generated data utility. Our approach leverages generative adversarial nets to generate data, combined with private aggregation among different discriminators to ensure strong privacy guarantees. Compared to existing approaches, G-PATE significantly improves the use of privacy budgets. In particular, we train a student data generator with an ensemble of teacher discriminators and propose a novel private gradient aggregation mechanism to ensure differential privacy on all information that flows from teacher discriminators to the student generator. In addition, with random projection and gradient discretization, the proposed gradient aggregation mechanism is able to effectively deal with high-dimensional gradient vectors. Theoretically, we prove that G-PATE ensures differential privacy for the data generator. Empirically, we demonstrate the superiority of G-PATE over prior work through extensive experiments. We show that G-PATE is the first work being able to generate high-dimensional image data with high data utility under limited privacy budgets ($ε\le 1$). Our code is available at https://github.com/AI-secure/G-PATE.

George Gunter, Huichen Li, Avesta Hojjati et al.

We demonstrate that a supply-chain level compromise of the adaptive cruise control (ACC) capability on equipped vehicles can be used to significantly degrade system level performance of current day mixed-autonomy freeway networks. Via a simple threat model which causes random deceleration attacks (RDAs), compromised vehicles create congestion waves in the traffic that decrease average speed and network throughput. We use a detailed and realistic traffic simulation environment to quantify the impacts of the attack on a model of a real high-volume freeway in the United States. We find that the effect of the attack depends both on the level of underlying traffic congestion, and what percentage of ACC vehicles can be compromised. In moderate congestion regimes the attack can degrade mean commuter speed by over 7%. In high density regimes overall network throughput can be reduced by up to 3%. And, in moderate to high congestion regimes, it can cost commuters on the network over 300 USD/km hr. All of these results motivate that the proposed attack is able to significantly degrade performance of the traffic network. We also develop an anomaly detection technique that uses GPS traces on vehicles to identify malicious/compromised vehicles. We employ this technique on data from the simulation experiments and find that it is unable to identify compromised ACCs compared to benign/normal drivers. That is, these attacks are stealthy to detection. Stronger attacks can be accurately labeled as malicious, motivating that there is a limit to how impactful attacks can be before they are no longer stealthy. Finally, we experimentally execute the attack on a real and commercially available ACC vehicle, demonstrating the possible real world feasibility of an RDA.

Hyun Bin Lee, Tushar M. Jois, Christopher W. Fletcher et al.

Users can improve the security of remote communications by using Trusted Execution Environments (TEEs) to protect against direct introspection and tampering of sensitive data. This can even be done with applications coded in high-level languages with complex programming stacks such as R, Python, and Ruby. However, this creates a trade-off between programming convenience versus the risk of attacks using microarchitectural side channels. In this paper, we argue that it is possible to address this problem for important applications by instrumenting a complex programming environment (like R) to produce a Data-Oblivious Transcript (DOT) that is explicitly designed to support computation that excludes side channels. Such a transcript is then evaluated on a Trusted Execution Environment (TEE) containing the sensitive data using a small trusted computing base called the Data-Oblivious Virtual Environment (DOVE). To motivate the problem, we demonstrate a number of subtle side-channel vulnerabilities in the R language. We then provide an illustrative design and implementation of DOVE for R, creating the first side-channel resistant R programming stack. We demonstrate that the two-phase architecture provided by DOT generation and DOVE evaluation can provide practical support for complex programming languages with usable performance and high security assurances against side channels.

Hsiao-Ying Huang, Soteris Demetriou, Rini Banerjee et al.

Despite widespread use of smartphones, there is no measurement standard targeted at smartphone security behaviors. In this paper we translate a well-known cybersecurity behavioral scale into the smartphone domain and show that we can improve on this translation by following an established psychometrics approach surveying 1011 participants. We design a new 14-item Smartphone Security Behavioral Scale (SSBS) exhibiting high reliability and good fit to a two-component behavioural model based on technical versus social protection strategies. We then demonstrate how SSBS can be applied to measure the influence of mental health issues on smartphone security behavior intentions. We found significant correlations that predict SSBS profiles from three types of MHIs. Conversely, we are able to predict presence of MHIs using SSBS profiles.We obtain prediction AUCs of 72.1% for Internet addiction,75.8% for depression and 66.2% for insomnia.

Xiaojun Xu, Qi Wang, Huichen Li et al.

In machine learning Trojan attacks, an adversary trains a corrupted model that obtains good performance on normal data but behaves maliciously on data samples with certain trigger patterns. Several approaches have been proposed to detect such attacks, but they make undesirable assumptions about the attack strategies or require direct access to the trained models, which restricts their utility in practice. This paper addresses these challenges by introducing a Meta Neural Trojan Detection (MNTD) pipeline that does not make assumptions on the attack strategies and only needs black-box access to models. The strategy is to train a meta-classifier that predicts whether a given target model is Trojaned. To train the meta-model without knowledge of the attack strategy, we introduce a technique called jumbo learning that samples a set of Trojaned models following a general distribution. We then dynamically optimize a query set together with the meta-classifier to distinguish between Trojaned and benign models. We evaluate MNTD with experiments on vision, speech, tabular data and natural language text datasets, and against different Trojan attacks such as data poisoning attack, model manipulation attack, and latent attack. We show that MNTD achieves 97% detection AUC score and significantly outperforms existing detection approaches. In addition, MNTD generalizes well and achieves high detection performance against unforeseen attacks. We also propose a robust MNTD pipeline which achieves 90% detection AUC even when the attacker aims to evade the detection with full knowledge of the system.

Michael J. May, Kevin D. Lux, Carl A. Gunter

Web services offer an opportunity to redesign a variety of older systems to exploit the advantages of a flexible, extensible, secure set of standards. In this work we revisit WSEmail, a system proposed over ten years ago to improve email by redesigning it as a family of web services. WSEmail offers an alternative vision of how instant messaging and email services could have evolved, offering security, extensibility, and openness in a distributed environment instead of the hardened walled gardens that today's rich messaging systems have become. WSEmail's architecture, especially its automatic plug-in download feature allows for rich extensions without changing the base protocol or libraries. We demonstrate WSEmail's flexibility using three business use cases: secure channel instant messaging, business workflows with routed forms, and on-demand attachments. Since increased flexibility often mitigates against security and performance, we designed WSEmail with security in mind and formally proved the security of one of its core protocols (on-demand attachments) using the TulaFale and ProVerif automated proof tools. We provide performance measurements for WSEmail functions in a prototype we implemented using .NET. Our experiments show a latency of about a quarter of a second per transaction under load.

Yunhui Long, Vincent Bindschaedler, Lei Wang et al.

Membership Inference Attack (MIA) determines the presence of a record in a machine learning model's training data by querying the model. Prior work has shown that the attack is feasible when the model is overfitted to its training data or when the adversary controls the training algorithm. However, when the model is not overfitted and the adversary does not control the training algorithm, the threat is not well understood. In this paper, we report a study that discovers overfitting to be a sufficient but not a necessary condition for an MIA to succeed. More specifically, we demonstrate that even a well-generalized model contains vulnerable instances subject to a new generalized MIA (GMIA). In GMIA, we use novel techniques for selecting vulnerable instances and detecting their subtle influences ignored by overfitting metrics. Specifically, we successfully identify individual records with high precision in real-world datasets by querying black-box machine learning models. Further we show that a vulnerable record can even be indirectly attacked by querying other related records and existing generalization techniques are found to be less effective in protecting the vulnerable instances. Our findings sharpen the understanding of the fundamental cause of the problem: the unique influences the training instance may have on the model.

Xuejing Yuan, Yuxuan Chen, Yue Zhao et al.

The popularity of ASR (automatic speech recognition) systems, like Google Voice, Cortana, brings in security concerns, as demonstrated by recent attacks. The impacts of such threats, however, are less clear, since they are either less stealthy (producing noise-like voice commands) or requiring the physical presence of an attack device (using ultrasound). In this paper, we demonstrate that not only are more practical and surreptitious attacks feasible but they can even be automatically constructed. Specifically, we find that the voice commands can be stealthily embedded into songs, which, when played, can effectively control the target system through ASR without being noticed. For this purpose, we developed novel techniques that address a key technical challenge: integrating the commands into a song in a way that can be effectively recognized by ASR through the air, in the presence of background noise, while not being detected by a human listener. Our research shows that this can be done automatically against real world ASR applications. We also demonstrate that such CommanderSongs can be spread through Internet (e.g., YouTube) and radio, potentially affecting millions of ASR users. We further present a new mitigation technique that controls this threat.

Yunhui Long, Vincent Bindschaedler, Carl A. Gunter

Machine learning models are increasingly made available to the masses through public query interfaces. Recent academic work has demonstrated that malicious users who can query such models are able to infer sensitive information about records within the training data. Differential privacy can thwart such attacks, but not all models can be readily trained to achieve this guarantee or to achieve it with acceptable utility loss. As a result, if a model is trained without differential privacy guarantee, little is known or can be said about the privacy risk of releasing it. In this work, we investigate and analyze membership attacks to understand why and how they succeed. Based on this understanding, we propose Differential Training Privacy (DTP), an empirical metric to estimate the privacy risk of publishing a classier when methods such as differential privacy cannot be applied. DTP is a measure of a classier with respect to its training dataset, and we show that calculating DTP is efficient in many practical cases. We empirically validate DTP using state-of-the-art machine learning models such as neural networks trained on real-world datasets. Our results show that DTP is highly predictive of the success of membership attacks and therefore reducing DTP also reduces the privacy risk. We advocate for DTP to be used as part of the decision-making process when considering publishing a classifier. To this end, we also suggest adopting the DTP-1 hypothesis: if a classifier has a DTP value above 1, it should not be published.

Vincent Bindschaedler, Reza Shokri, Carl A. Gunter

Releasing full data records is one of the most challenging problems in data privacy. On the one hand, many of the popular techniques such as data de-identification are problematic because of their dependence on the background knowledge of adversaries. On the other hand, rigorous methods such as the exponential mechanism for differential privacy are often computationally impractical to use for releasing high dimensional data or cannot preserve high utility of original data due to their extensive data perturbation. This paper presents a criterion called plausible deniability that provides a formal privacy guarantee, notably for releasing sensitive datasets: an output record can be released only if a certain amount of input records are indistinguishable, up to a privacy parameter. This notion does not depend on the background knowledge of an adversary. Also, it can efficiently be checked by privacy tests. We present mechanisms to generate synthetic datasets with similar statistical properties to the input data and the same format. We study this technique both theoretically and experimentally. A key theoretical result shows that, with proper randomization, the plausible deniability mechanism generates differentially private synthetic data. We demonstrate the efficiency of this generative technique on a large dataset; it is shown to preserve the utility of original data with respect to various statistical analysis and machine learning measures.

Wenhao Wang, Guoxing Chen, Xiaorui Pan et al.

Side-channel risks of Intel's SGX have recently attracted great attention. Under the spotlight is the newly discovered page-fault attack, in which an OS-level adversary induces page faults to observe the page-level access patterns of a protected process running in an SGX enclave. With almost all proposed defense focusing on this attack, little is known about whether such efforts indeed raise the bar for the adversary, whether a simple variation of the attack renders all protection ineffective, not to mention an in-depth understanding of other attack surfaces in the SGX system. In the paper, we report the first step toward systematic analyses of side-channel threats that SGX faces, focusing on the risks associated with its memory management. Our research identifies 8 potential attack vectors, ranging from TLB to DRAM modules. More importantly, we highlight the common misunderstandings about SGX memory side channels, demonstrating that high frequent AEXs can be avoided when recovering EdDSA secret key through a new page channel and fine-grained monitoring of enclave programs (at the level of 64B) can be done through combining both cache and cross-enclave DRAM channels. Our findings reveal the gap between the ongoing security research on SGX and its side-channel weaknesses, redefine the side-channel threat model for secure enclaves, and can provoke a discussion on when to use such a system and how to use it securely.

Nan Zhang, Soteris Demetriou, Xianghang Mi et al.

Inspired by the boom of the consumer IoT market, many device manufacturers, start-up companies and technology giants have jumped into the space. Unfortunately, the exciting utility and rapid marketization of IoT, come at the expense of privacy and security. Industry reports and academic work have revealed many attacks on IoT systems, resulting in privacy leakage, property loss and large-scale availability problems. To mitigate such threats, a few solutions have been proposed. However, it is still less clear what are the impacts they can have on the IoT ecosystem. In this work, we aim to perform a comprehensive study on reported attacks and defenses in the realm of IoT aiming to find out what we know, where the current studies fall short and how to move forward. To this end, we first build a toolkit that searches through massive amount of online data using semantic analysis to identify over 3000 IoT-related articles. Further, by clustering such collected data using machine learning technologies, we are able to compare academic views with the findings from industry and other sources, in an attempt to understand the gaps between them, the trend of the IoT security risks and new problems that need further attention. We systemize this process, by proposing a taxonomy for the IoT ecosystem and organizing IoT security into five problem areas. We use this taxonomy as a beacon to assess each IoT work across a number of properties we define. Our assessment reveals that relevant security and privacy problems are far from solved. We discuss how each proposed solution can be applied to a problem area and highlight their strengths, assumptions and constraints. We stress the need for a security framework for IoT vendors and discuss the trend of shifting security liability to external or centralized entities. We also identify open research problems and provide suggestions towards a secure IoT ecosystem.

Muhammad Naveed, Erman Ayday, Ellen W. Clayton et al.

Genome sequencing technology has advanced at a rapid pace and it is now possible to generate highly-detailed genotypes inexpensively. The collection and analysis of such data has the potential to support various applications, including personalized medical services. While the benefits of the genomics revolution are trumpeted by the biomedical community, the increased availability of such data has major implications for personal privacy; notably because the genome has certain essential features, which include (but are not limited to) (i) an association with traits and certain diseases, (ii) identification capability (e.g., forensics), and (iii) revelation of family relationships. Moreover, direct-to-consumer DNA testing increases the likelihood that genome data will be made available in less regulated environments, such as the Internet and for-profit companies. The problem of genome data privacy thus resides at the crossroads of computer science, medicine, and public policy. While the computer scientists have addressed data privacy for various data types, there has been less attention dedicated to genomic data. Thus, the goal of this paper is to provide a systematization of knowledge for the computer science community. In doing so, we address some of the (sometimes erroneous) beliefs of this field and we report on a survey we conducted about genome data privacy with biomedical specialists. Then, after characterizing the genome privacy problem, we review the state-of-the-art regarding privacy attacks on genomic data and strategies for mitigating such attacks, as well as contextualizing these attacks from the perspective of medicine and public policy. This paper concludes with an enumeration of the challenges for genome data privacy and presents a framework to systematize the analysis of threats and the design of countermeasures as the field moves forward.

Michael LeMay, Carl A. Gunter

Mobile devices are in roles where the integrity and confidentiality of their apps and data are of paramount importance. They usually contain a System-on-Chip (SoC), which integrates microprocessors and peripheral Intellectual Property (IP) connected by a Network-on-Chip (NoC). Malicious IP or software could compromise critical data. Some types of attacks can be blocked by controlling data transfers on the NoC using Memory Management Units (MMUs) and other access control mechanisms. However, commodity processors do not provide strong assurances regarding the correctness of such mechanisms, and it is challenging to verify that all access control mechanisms in the system are correctly configured. We propose a NoC Firewall (NoCF) that provides a single locus of control and is amenable to formal analysis. We demonstrate an initial analysis of its ability to resist malformed NoC commands, which we believe is the first effort to detect vulnerabilities that arise from NoC protocol violations perpetrated by erroneous or malicious IP.