Biblio

We consider the problem of privacy-sensitive anomaly detection - screening to detect individuals, behaviors, areas, or data samples of high interest. What defines an anomaly is context-specific; for example, a spoofed rather than genuine user attempting to log in to a web site, a fraudulent credit card transaction, or a suspicious traveler in an airport. The unifying assumption is that the number of anomalous points is quite small with respect to the population, so that deep screening of all individual data points would potentially be time-intensive, costly, and unnecessarily invasive of privacy. Such privacy violations can raise concerns due sensitive nature of data being used, raise fears about violations of data use agreements, and make people uncomfortable with anomaly detection methods. Anomaly detection is well studied, but methods to provide anomaly detection along with privacy are less well studied. Our overall goal in this research is to provide a framework for identifying anomalous data while guaranteeing quantifiable privacy in a rigorous sense. Once identified, such anomalies could warrant further data collection and investigation, depending on the context and relevant policies. In this research, we focus on privacy protection during the deployment of anomaly detection. Our main contribution is a differentially private access mechanism for finding anomalies using a search algorithm based on adaptive noisy group testing. To achieve this, we take as our starting point the notion of group testing [1], which was most famously used to screen US military draftees for syphilis during World War II. In group testing, individuals are tested in groups to limit the number of tests. Using multiple rounds of screenings, a small number of positive individuals can be detected very efficiently. Group testing has the added benefit of providing privacy to individuals through plausible deniability - since the group tests use aggregate data, individual contributions to the test are masked by the group. We follow on these concepts by demonstrating a search model utilizing adaptive queries on aggregated group data. Our work takes the first steps toward strengthening and formalizing these privacy concepts by achieving differential privacy [2]. Differential privacy is a statistical measure of disclosure risk that captures the intuition that an individual's privacy is protected if the results of a computation have at most a very small and quantifiable dependence on that individual's data. In the last decade, there hpractical adoption underway by high-profile companies such as Apple, Google, and Uber. In order to make differential privacy meaningful in the context of a task that seeks to specifically identify some (anomalous) individuals, we introduce the notion of anomaly-restricted differential privacy. Using ideas from information theory, we show that noise can be added to group query results in a way that provides differential privacy for non-anomalous individuals and still enables efficient and accurate detection of the anomalous individuals. Our method ensures that using differentially private aggregation of groups of points, providing privacy to individuals within the group while refining the group selection to the point that we can probabilistically narrow attention to a small numbers of individuals or samples for further attention. To summarize: We introduce a new notion of anomaly-restriction differential privacy, which may be of independent interest. We provide a noisy group-based search algorithm that satisfies the anomaly-restricted differential privacy definition. We provide both theoretical and empirical analysis of our noisy search algorithm, showing that it performs well in some cases, and exhibits the usual privacy/accuracy tradeoff of differentially private mechanisms. Potential anomaly detection applications for our work might include spatial search for outliers: this would rely on new sensing technologies that can perform queries in aggregate to reveal and isolate anomalous outliers. For example, this could lead to privacy-sensitive methods for searching for outlying cell phone activity patterns or Internet activity patterns in a geographic location.

In this paper, we present a differential privacy preserving approach, which extracts personality-based knowledge to serve privacy guarantee data analysis on personal sensitive data. Based on the approach, we further implement an end-to-end privacy guarantee system, KaPPA, to provide researchers iterative data analysis on sensitive data. The key challenge for differential privacy is determining a reasonable amount of privacy budget to balance privacy preserving and data utility. Most of the previous work applies unified privacy budget to all individual data, which leads to insufficient privacy protection for some individuals while over-protecting others. In KaPPA, the proposed personality-based privacy preserving approach automatically calculates privacy budget for each individual. Our experimental evaluations show a significant trade-off of sufficient privacy protection and data utility.

Lately, mobile operators experience threats from SMS grey routes which are used by fraudsters to evade SMS fees and to deny them millions in revenues. But more serious are the threats to the user's security and privacy and consequently the operator's reputation. Therefore, it is crucial for operators to have adequate solutions to protect both their network and their customers against this kind of fraud. Unfortunately, so far there is no sufficiently efficient countermeasure against grey routes. This paper proposes a near real time SMS grey traffic detection which makes use of Counting Bloom Filters combined with blacklist and whitelist to detect SMS grey traffic on the fly and to block them. The proposed detection has been implemented and proved to be quite efficient. The paper provides also comprehensive explanation of SMS grey routes and the challenges in their detection. The implementation and verification are also described thoroughly.

In machine learning, feature engineering has been a pivotal stage in building a high-quality predictor. Particularly, this work explores the multiple Kernel Discriminant Component Analysis (mKDCA) feature-map and its variants. However, seeking the right subset of kernels for mKDCA feature-map can be challenging. Therefore, we consider the problem of kernel selection, and propose an algorithm based on Differential Mutual Information (DMI) and incremental forward search. DMI serves as an effective metric for selecting kernels, as is theoretically supported by mutual information and Fisher's discriminant analysis. On the other hand, incremental forward search plays a role in removing redundancy among kernels. Finally, we illustrate the potential of the method via an application in privacy-aware classification, and show on three mobile-sensing datasets that selecting an effective set of kernels for mKDCA feature-maps can enhance the utility classification performance, while successfully preserve the data privacy. Specifically, the results show that the proposed DMI forward search method can perform better than the state-of-the-art, and, with much smaller computational cost, can perform as well as the optimal, yet computationally expensive, exhaustive search.

In this paper, we focus on developing a novel mechanism to preserve differential privacy in deep neural networks, such that: (1) The privacy budget consumption is totally independent of the number of training steps; (2) It has the ability to adaptively inject noise into features based on the contribution of each to the output; and (3) It could be applied in a variety of different deep neural networks. To achieve this, we figure out a way to perturb affine transformations of neurons, and loss functions used in deep neural networks. In addition, our mechanism intentionally adds "more noise" into features which are "less relevant" to the model output, and vice-versa. Our theoretical analysis further derives the sensitivities and error bounds of our mechanism. Rigorous experiments conducted on MNIST and CIFAR-10 datasets show that our mechanism is highly effective and outperforms existing solutions.

We propose a privacy-preserving framework for learning visual classifiers by leveraging distributed private image data. This framework is designed to aggregate multiple classifiers updated locally using private data and to ensure that no private information about the data is exposed during and after its learning procedure. We utilize a homomorphic cryptosystem that can aggregate the local classifiers while they are encrypted and thus kept secret. To overcome the high computational cost of homomorphic encryption of high-dimensional classifiers, we (1) impose sparsity constraints on local classifier updates and (2) propose a novel efficient encryption scheme named doublypermuted homomorphic encryption (DPHE) which is tailored to sparse high-dimensional data. DPHE (i) decomposes sparse data into its constituent non-zero values and their corresponding support indices, (ii) applies homomorphic encryption only to the non-zero values, and (iii) employs double permutations on the support indices to make them secret. Our experimental evaluation on several public datasets shows that the proposed approach achieves comparable performance against state-of-the-art visual recognition methods while preserving privacy and significantly outperforms other privacy-preserving methods.

Safety-critical system engineering and traditional safety analyses have for decades been focused on problems caused by natural or accidental phenomena. Security analyses, on the other hand, focus on preventing intentional, malicious acts that reduce system availability, degrade user privacy, or enable unauthorized access. In the context of safety-critical systems, safety and security are intertwined, e.g., injecting malicious control commands may lead to system actuation that causes harm. Despite this intertwining, safety and security concerns have traditionally been designed and analyzed independently of one another, and examined in very different ways. In this work we examine a new hazard analysis technique—Systematic Analysis of Faults and Errors (SAFE)—and its deep integration of safety and security concerns. This is achieved by explicitly incorporating a semantic framework of error "effects" that unifies an adversary model long used in security contexts with a fault/error categorization that aligns with previous approaches to hazard analysis. This categorization enables analysts to separate the immediate, component-level effects of errors from their cause or precise deviation from specification. This paper details SAFE's integrated handling of safety and security through a) a methodology grounded in—and adaptable to—different approaches from the literature, b) explicit documentation of system assumptions which are implicit in other analyses, and c) increasing the tractability of analyzing modern, complex, component-based software-driven systems. We then discuss how SAFE's approach supports the long-term goals of of increased compositionality and formalization of safety/security analysis. 

In recent works, numerous physical-layer security systems have been proposed as alternatives to classic cryptography. Such systems aim to use the intrinsic properties of radio signals and the wireless medium to provide confidentiality and authentication to wireless devices. However, fundamental vulnerabilities are often discovered in these systems shortly after their inception. We therefore challenge the assumptions made by existing physical-layer security systems, and postulate that weaker assumptions are needed in order to adapt for practical scenarios. We also argue that if no computational advantage over an adversary can be ensured, secure communication cannot be realistically achieved.

Multileaved comparison methods generalize interleaved comparison methods to provide a scalable approach for comparing ranking systems based on regular user interactions. Such methods enable the increasingly rapid research and development of search engines. However, existing multileaved comparison methods that provide reliable outcomes do so by degrading the user experience during evaluation. Conversely, current multileaved comparison methods that maintain the user experience cannot guarantee correctness. Our contribution is two-fold. First, we propose a theoretical framework for systematically comparing multileaved comparison methods using the notions of considerateness, which concerns maintaining the user experience, and fidelity, which concerns reliable correct outcomes. Second, we introduce a novel multileaved comparison method, Pairwise Preference Multileaving (PPM), that performs comparisons based on document-pair preferences, and prove that it is considerate and has fidelity. We show empirically that, compared to previous multileaved comparison methods, PPM is more sensitive to user preferences and scalable with the number of rankers being compared.

The vision of smart environments, systems, and services is driven by the development of the Internet of Things (IoT). IoT devices produce large amounts of data and this data is used to make critical decisions in many systems. The data produced by these devices has to satisfy various security related requirements in order to be useful in practical scenarios. One of these requirements is data provenance which allows a user to trust the data regarding its origin and location. The low cost of many IoT devices and the fact that they may be deployed in unprotected spaces requires security protocols to be efficient and secure against physical attacks. This paper proposes a light-weight protocol for data provenance in the IoT. The proposed protocol uses physical unclonable functions (PUFs) to provide physical security and uniquely identify an IoT device. Moreover, wireless channel characteristics are used to uniquely identify a wireless link between an IoT device and a server/user. A brief security and performance analysis are presented to give a preliminary validation of the protocol.

Cyber risk management largely reduces to a race for information between defenders of ICT systems and attackers. Defenders can gain advantage in this race by sharing cyber risk information with each other. Yet, they often exchange less information than is socially desirable, because sharing decisions are guided by selfish rather than altruistic reasons. A growing line of research studies these strategic aspects that drive defenders’ sharing decisions. The present survey systematizes these works in a novel framework. It provides a consolidated understanding of defenders’ strategies to privately or publicly share information and enables us to distill trends in the literature and identify future research directions. We reveal that many theoretical works assume cyber risk information sharing to be beneficial, while empirical validations are often missing.

In this paper, a novel secure information exchange scheme has been proposed for MIMO vehicular ad hoc networks (VANETs) through physical layer approach. In the scheme, a group of On Board Units (OBUs) exchange information with help of one Road Side Unit (RSU). By utilizing the key signal processing technique, i.e., Direction Rotation Alignment technique, the information to be exchanged of the two neighbor OBUs are aligned into a same direction to form summed signal at RSU or external eavesdroppers. With such summed signal, the RSU or the eavesdropper cannot recover the individual information from the OBUs. By regulating the transmission rate for each OBU, the information theoretic security could be achieved. The secrecy sum-rates of the proposed scheme are analyzed following the scheme. Finally, the numerical results are conducted to demonstrate the theoretical analysis.

A wireless sensor network (WSN) is composed of sensor nodes and a base station. In WSNs, constructing an efficient key-sharing scheme to ensure a secure communication is important. In this paper, we propose a new key-sharing scheme for groups, which shares a group key in a single broadcast without being dependent on the number of nodes. This scheme is based on geometric characteristics and has information-theoretic security in the analysis of transmitted data. We compared our scheme with conventional schemes in terms of communication traffic, computational complexity, flexibility, and security, and the results showed that our scheme is suitable for an Internet-of-Things (IoT) network.

In Wyner wiretap II model of communication, Alice and Bob are connected by a channel that can be eavesdropped by an adversary with unlimited computation who can select a fraction of communication to view, and the goal is to provide perfect information theoretic security. Information theoretic security is increasingly important because of the threat of quantum computers that can effectively break algorithms and protocols that are used in today's public key infrastructure. We consider interactive protocols for wiretap II channel with active adversary who can eavesdrop and add adversarial noise to the eavesdropped part of the codeword. These channels capture wireless setting where malicious eavesdroppers at reception distance of the transmitter can eavesdrop the communication and introduce jamming signal to the channel. We derive a new upperbound R ≤ 1 - ρ for the rate of interactive protocols over two-way wiretap II channel with active adversaries, and construct a perfectly secure protocol family with achievable rate 1 - 2ρ + ρ2. This is strictly higher than the rate of the best one round protocol which is 1 - 2ρ, hence showing that interaction improves rate. We also prove that even with interaction, reliable communication is possible only if ρ \textbackslashtextless; 1/2. An interesting aspect of this work is that our bounds will also hold in network setting when two nodes are connected by n paths, a ρ of which is corrupted by the adversary. We discuss our results, give their relations to the other works, and propose directions for future work.

In fiber-optic communication networks, research on data security at lower layers of the protocol stack and in particular at the physical layer by means of information-theoretic concepts is only in the beginning. Nevertheless, it has recently attracted quite some attention as it holds the promise of providing unconditional, perfect security without the need for secret key exchanges. In this paper, we analyze some important constraints that such concepts put on a potential implementation of physical-layer security. We review the fundamentals of physical-layer security on the basis of the commonly used AWGN wiretap channel model. For such channel model we summarize the security metrics which are typically used in information theory and in particular recall that, for secure communication over the AWGN channel, the legitimate receiver needs an SNR advantage over the eavesdropper. Next, we relate the information theoretic metrics to physically measurable quantities in optical communications engineering, namely optical signal-to-noise ratio (OSNR) and bit-error ratio (BER), and translate the information-theoretic wiretap scenario to a simple real-world point-to-point optical transmission link in which part of the light is wiretapped using a bend coupler. We investigate the achievable OSNR advantage under realistic assumptions for fiber loss, tap ratio, and noise budget and find that secure transmission is limited to a distance of a few tens of kilometers in this case. The maximum secure transmission distance decreases with an increasing tap ratio chosen by the eavesdropper. This can be only counteracted by monitoring the link loss towards the legitimate receiver which would force the eavesdropper to choose small tap ratios in order to remain undetected. In an outlook towards further research directions we identify information-theoretic approaches which could potentially allow to realize physical-layer security in more generalized scenarios or over longer distances.

We report on our research on proving the security of multi-party cryptographic protocols using the EASYCRYPT proof assistant. We work in the computational model using the sequence of games approach, and define honest-butcurious (semi-honest) security using a variation of the real/ideal paradigm in which, for each protocol party, an adversary chooses protocol inputs in an attempt to distinguish the party's real and ideal games. Our proofs are information-theoretic, instead of being based on complexity theory and computational assumptions. We employ oracles (e.g., random oracles for hashing) whose encapsulated states depend on dynamically-made, nonprogrammable random choices. By limiting an adversary's oracle use, one may obtain concrete upper bounds on the distances between a party's real and ideal games that are expressed in terms of game parameters. Furthermore, our proofs work for adaptive adversaries, ones that, when choosing the value of a protocol input, may condition this choice on their current protocol view and oracle knowledge. We provide an analysis in EASYCRYPT of a three party private count retrieval protocol. We emphasize the lessons learned from completing this proof.

The rapid digitalisation of the hospitality industry over recent years has brought forth many new points of attack for consideration. The hasty implementation of these systems has created a reality in which businesses are using the technical solutions, but employees have very little awareness when it comes to the threats and implications that they might present. This gap in awareness is further compounded by the existence of preestablished, often rigid, cultures that drive how hospitality businesses operate. Potential attackers are recognising this and the last two years have seen a huge increase in cyber-attacks within the sector.Attempts at addressing the increasing threats have taken the form of technical solutions such as encryption, access control, CCTV, etc. However, a high majority of security breaches can be directly attributed to human error. It is therefore necessary that measures for addressing the rising trend of cyber-attacks go beyond just providing technical solutions and make provision for educating employees about how to address the human elements of security. Inculcating security awareness amongst hospitality employees will provide a foundation upon which a culture of security can be created to promote the seamless and secured interaction of hotel users and technology.One way that the hospitality industry has tried to solve the awareness issue is through their current paper-based training. This is unengaging, expensive and presents limited ways to deploy, monitor and evaluate the impact and effectiveness of the content. This leads to cycles of constant training, making it very hard to initiate awareness, particularly within those on minimum waged, short-term job roles.This paper presents a structured approach for eliciting industry requirement for developing and implementing an immersive Cyber Security Awareness learning platform. It used a series of over 40 interviews and threat analysis of the hospitality industry to identify the requirements fo- designing and implementing cyber security program which encourage engagement through a cycle of reward and recognition. In particular, the need for the use of gamification elements to provide an engaging but gentle way of educating those with little or no desire to learn was identified and implemented. Also presented is a method for guiding and monitoring the impact of their employee's progress through the learning management system whilst monitoring the levels of engagement and positive impact the training is having on the business.

End users are prone to insecure cyber behavior that may lead them to compromise the integrity, availability or confidentiality of their computer systems. For instance, replying to a phishing email may compromise an end user's login credentials. Identifying tendency toward insecure cyber behavior is critically important to improve cyber security posture and thesis of this paper is that the susceptibility of end-users to be a victim of a cyber-attack may be predicted using personality traits such as trait anxiety and callousness. This paper presents an easily configurable, script-based software tool to explore the relationships between the personality traits and insecure cyber behaviors of end users. The software utilizes well-established cognitive methods (such as dot probe) to identify a number of personality traits for a user and further allows researchers to design and conduct experiments through customizable scripting to study the endusers' insecure cyber behaviors. The software also collects fine-grained data on users for analysis.

Computing systems today have a large number of security configuration settings that enforce security properties. However, vulnerabilities and incorrect configuration increase the potential for attacks. Provable verification and simulation tools have been introduced to eliminate configuration conflicts and weaknesses, which can increase system robustness against attacks. Most of these tools require special knowledge in formal methods and precise specification for requirements in special languages, in addition to their excessive need for computing resources. Video games have been utilized by researchers to make educational software more attractive and engaging. Publishing these games for crowdsourcing can also stimulate competition between players and increase the game educational value. In this paper we introduce a game interface, called NetMaze, that represents the network configuration verification problem as a video game and allows for attack analysis. We aim to make the security analysis and hardening usable and accurately achievable, using the power of video games and the wisdom of crowdsourcing. Players can easily discover weaknesses in network configuration and investigate new attack scenarios. In addition, the gameplay scenarios can also be used to analyze and learn attack attribution considering human factors. In this paper, we present a provable mapping from the network configuration to 3D game objects.