Privacy Attacks Research Articles

Modelling Memory for Individual Re-identification in Decentralised Mobile Contact Tracing Applications

In 2020 the coronavirus outbreak changed the lives of people worldwide. After an initial time period in which it was unclear how to battle the virus, social distancing has been recognised globally as an effective method to mitigate the disease spread. This called for technological tools such as Mobile Contact Tracing Applications (MCTA), which are used to digitally trace contacts among people, and in case a positive case is found, people with the application installed which had been in contact will be notified. De-centralised MCTA may suffer from a novel kind of privacy attack, based on the memory of the human beings, which upon notification of the application can identify who is the positive individual responsible for the notification. Our results show that it is indeed possible to identify positive people among the group of contacts of a human being, and this is even easier when the sociability of the positive individual is low. In practice, our simulation results show that identification can be made with an accuracy of more than 90% depending on the scenario. We also provide three mitigation strategies which can be implemented in de-centralised MCTA and analyse which of the three are more effective in limiting this novel kind of attack.

When Machine Learning Meets Privacy

The newly emerged machine learning (e.g., deep learning) methods have become a strong driving force to revolutionize a wide range of industries, such as smart healthcare, financial technology, and surveillance systems. Meanwhile, privacy has emerged as a big concern in this machine learning-based artificial intelligence era. It is important to note that the problem of privacy preservation in the context of machine learning is quite different from that in traditional data privacy protection, as machine learning can act as both friend and foe. Currently, the work on the preservation of privacy and machine learning are still in an infancy stage, as most existing solutions only focus on privacy problems during the machine learning process. Therefore, a comprehensive study on the privacy preservation problems and machine learning is required. This article surveys the state of the art in privacy issues and solutions for machine learning. The survey covers three categories of interactions between privacy and machine learning: (i) private machine learning, (ii) machine learning-aided privacy protection, and (iii) machine learning-based privacy attack and corresponding protection schemes. The current research progress in each category is reviewed and the key challenges are identified. Finally, based on our in-depth analysis of the area of privacy and machine learning, we point out future research directions in this field.

Online Algorithm for Differentially Private Genome-wide Association Studies

Digitization of healthcare records contributed to a large volume of functional scientific data that can help researchers to understand the behaviour of many diseases. However, the privacy implications of this data, particularly genomics data, have surfaced recently as the collection, dissemination, and analysis of human genomics data is highly sensitive. There have been multiple privacy attacks relying on the uniqueness of the human genome that reveals a participant or a certain group’s presence in a dataset. Therefore, the current data sharing policies have ruled out any public dissemination and adopted precautionary measures prior to genomics data release, which hinders timely scientific innovation. In this article, we investigate an approach that only releases the statistics from genomic data rather than the whole dataset and propose a generalized Differentially Private mechanism for Genome-wide Association Studies (GWAS). Our method provides a quantifiable privacy guarantee that adds noise to the intermediate outputs but ensures satisfactory accuracy of the private results. Furthermore, the proposed method offers multiple adjustable parameters that the data owners can set based on the optimal privacy requirements. These variables are presented as equalizers that balance between the privacy and utility of the GWAS. The method also incorporates Online Bin Packing technique [1], which further bounds the privacy loss linearly, growing according to the number of open bins and scales with the incoming queries. Finally, we implemented and benchmarked our approach using seven different GWAS studies to test the performance of the proposed methods. The experimental results demonstrate that for 1,000 arbitrary online queries, our algorithms are more than 80% accurate with reasonable privacy loss and exceed the state-of-the-art approaches on multiple studies (i.e., EigenStrat, LMM, TDT).

Privacy preservation of electronic health records with adversarial attacks identification in hybrid cloud

An increasing trend in healthcare organizations to outsource EHRs’ data to the cloud highlights new challenges regarding the privacy of given individuals. Healthcare organizations outsource their EHRs data in a hybrid cloud that elevates the problem of security and privacy in terms of EHRs’ access to an unlimited number of recipients in a hybrid cloud environment. In this paper, we investigated the need for a privacy-preserving access control model for the hybrid cloud. A comprehensive and exploratory analysis of privacy-preserving solutions with the help of taxonomy for cloud-based EHRs is described in this work. We have formally identified the existence of internal access control and external privacy disclosures in outsourcing system architecture for hybrid cloud. Then, we proposed a privacy-preserving XACML based access control model (PPX-AC) that supports fine-grained access control with the multipurpose utilization of EHRs alongside state-of-the-art privacy mechanism. Our proposed approach invalidates the identified security and privacy attacks. We have formally verified the proposed privacy-preserving XACML based access control model (PPX-AC) with the invalidation of identified privacy attacks using High-Level Petri Nets (HLPN). Moreover, property verification of the proposed model in SMT-lib and Z3 solver and implementation of the model proves its effectiveness in terms of privacy-aware EHRs access and multipurpose usage.

An efficient clustering-based anonymization scheme for privacy-preserving data collection in IoT based healthcare services

The healthcare services industry has seen a huge transformation since the prominent rise of the Internet of Things (IoT). IoT in healthcare services includes a large number of unified and interconnected sensors, and medical devices that generate and exchange sensitive information. Thus, an enormous amount of data is transmitted through the network which raises an alarming concern for the privacy of patient information. Therefore, privacy preserving data collection (PPDC) is on-demand to ensure the privacy of patient data. Several pieces of research on PPDC have been proposed recently. However, the research literatures have fallen short in privacy requirements and are prone to various privacy attacks. In this paper, we propose a novel privacy-preserving data collection scheme for IoT based healthcare services systems. A clustering-based anonymity model is utilized to develop an efficient privacy-preserving scheme to meet privacy requirements and to prevent healthcare IoT from various privacy attacks. We formulated the threat model as client-server-to-user to ensure privacy on both ends. On the client-side, a modified clustering-based k-anonymity model with α-deassociation is used to anonymize the data generated from the IoT nodes. The base-level privacy is then ensured through a bottom-up clustering method which generates clusters of records as per the privacy requirements. On the server-side, the cluster-combination method-UPGMA is utilized to reduce communication costs and to achieve a better level of privacy. The proposed scheme is efficient in tackling privacy attacks such as attribute disclosure, identity disclosure, membership disclosure, sensitivity attacks, similarity attacks, and skewness attacks. The effectiveness and efficiency of the proposed scheme are proven through theoretical and experimental analyses.

A robust privacy preserving approach for electronic health records using multiple dataset with multiple sensitive attributes

Privacy preserving data publishing of electronic health record (EHRs) for 1 to M datasets with multiple sensitive attributes (MSAs) is an interesting and challenging issue. There is always a trade-off between privacy and utility in data publishing. Most of the privacy-preserving models shows critical privacy disclosure issues and, hence, they are not robust in practical datasets. The k-anonymity model is a broadly used privacy model to analyze privacy disclosures, however, this model is only useful against identity disclosure. To address the limitations of k-anonymity, a group of privacy model extensions have been proposed in past years. It includes a p-sensitive k-anonymity model, a p+-sensitive k-anonymity model, and a balanced p+-sensitive k-anonymity model. However these privacy-preserving models are not sufficient to preserve the privacy of end-users in practical datasets. In this paper we have formalize the behavior of an adversary which perform identity and attribute disclosures on balanced p+-sensitive k-anonymity model with the help of adversarial scenarios. Since balanced p+-sensitive k-anonymity model is not sufficient for 1 to M with MSAs datasets privacy preservation. We propose an extended privacy model called “1: M MSA-(p, l)-diversity” for 1: M dataset with MSAs. We then perform formal modeling and verification of the proposed model using High-Level Petri Nets (HLPN) to confirm privacy attacks invalidation. Experimental results show that our proposed “1: M MSA-(p, l)-diversity model” is efficient and provide enhanced data utility of published data.

Privacy-preserving and communication-efficient federated learning in Internet of Things

Aimed at the privacy leakage caused by collecting data from numerous Internet of Things (IoT) devices for centralized training, a novel distributed learning framework, namely federated learning, came into being, where devices train models collaboratively while leaving their private datasets locally. Although many schemes have been proposed about federated learning, they are still short in communications and privacy due to the limited network bandwidth and advanced privacy attacks. To address these challenges, we develop PCFL, a privacy-preserving and communication-efficient scheme for federated learning in IoT. PCFL is composed of three key components: (1) gradient spatial sparsification where irrelevant local updates that deviate from the collaborative convergence tendency are prevented from being uploaded; (2) bidirectional compression where computation-less compression operators are used to quantize the gradients both at the device-side and server-side; and (3) privacy-preserving protocol which integrates secret sharing with lightweight homomorphic encryption to protect the data privacy and resist against various collusion scenarios. We analyze the correctness and privacy of our scheme, and carry out theoretical and experimental comparison on two real-world datasets. Results show that PCFL outperforms the state-of-the-art methods by more than 2× in terms of communication efficiency, along with high model accuracy and marginal decreases in convergence rate.

Secure Secret Sharing Enabled b-band Mini Vaults Bio-Cryptosystem for Vectorial Biometrics

Biometric Cryptosystems for secret binding such as fuzzy vault and fuzzy commitment are provable secure and offers a convenient way for secret management and protection. Despite numerous practical schemes have been reported, they are deficient in resisting several security and privacy attacks. In this paper, we propose a novel bio-cryptosystem that based on the three key ingredients namely Index of Maximum (IoM) hashing, $(m, k)$ ( m , k ) threshold secret sharing and b -band mini vaults notion. The IoM hashing is motivated from the ranking based Locality Sensitive Hashing theory meant for non-invertible transformation. On the other hand, the $(m, k)$ ( m , k ) threshold secret sharing scheme and the b -band mini vaults manage overcome inherent limitations of biometric cryptosystems when integrated with IoM hashing. The proposed scheme strikes the balance between performance and the privacy/security protection. Unlike fuzzy vault and fuzzy commitment, which primarily devised for unordered and binary biometrics, respectively, our scheme is tailored for feature vector-based biometrics (vectorial biometrics). Comprehensive experiments on fingerprint vectors that derived from several FVC fingerprint benchmarks and rigorous analysis demonstrate decent secret retrieval performance yet offer strong resilience against six major security and privacy attacks.

Efficient Known-Sample Attack for Distance-Preserving Hashing Biometric Template Protection Schemes

The rapid deployment of biometric authentication systems raises concern over user privacy and security. A biometric template protection scheme emerges as a solution to protect individual biometric templates stored in a database. Among all available protection schemes, a template protection scheme that relies on distance-preserving hashing has received much attention due to its simplicity and efficiency in offering privacy protection while archiving decent authentication performance. In this work, we introduce an efficient attack called known sample attack and demonstrate that most state-of-art template protection schemes that utilize distance-preserving hashing can be compromised in practice (within few seconds), especially when the output is significantly smaller than the original input sample size. These findings further motivated our subsequent work in proposing a secure authentication mechanism to resist such an attack with proper study over the distribution of the input samples. Furthermore, we conducted revocability, unlinkability analysis to demonstrate the satisfactory of general biometric template protection requirements; and showed the resistance of various security and privacy attacks, i.e., false acceptance attack, and attack via record multiplicity.

КЛАССИФИКАЦИЯ МЕХАНИЗМОВ АТАК И ИССЛЕДОВАНИЕ МЕТОДОВ ЗАЩИТЫ СИСТЕМ С ИСПОЛЬЗОВАНИЕМ АЛГОРИТМОВ МАШИННОГО ОБУЧЕНИЯ И ИСКУССТВЕННОГО ИНТЕЛЛЕКТА

This article provides a complete classification of attacks using artificial intelligence. Three main identified sections were considered: attacks on information systems and computer networks, attacks on artificial intelligence models (poisoning attacks, evasion attacks, extraction attacks, privacy attacks), attacks on human consciousness and opinion (all types of deepfake). In each of these sections, the mechanisms of attacks were identified and studied, in accordance with them, the methods of protection were set. In conclusion, a specific example of an attack using a pretrained model was analyzed and protected against it using the method of modifying the input data, namely, image compression in order to get rid of extraneous noise.

A Semantically Sensitive Privacy Protection Method for Trajectory Publishing

Trajectory data set is the indispensable foundation for constructing reliable Internet of Vehicles (IoV) service and location-based service (LBS), while it is likely to be abused by malicious attackers to infer user’s privacy. In this paper, we propose a trajectory protection method based on stop points obfuscation, which can confront various privacy attacks and preserve the semantic information to achieve adequate utility. Two strategies for stop point selection are designed, including category-distance priority method and Markov matrix method. Our new method was analyzed and evaluated on a real-world trajectory data set. The experiment result shows that our method can improve the utility of the data set and provide multi-level privacy protection.

Fingerprint Based Authentication Architecture for Accessing Multiple Cloud Computing Services using Single User Credential in IOT Environments

Cloud computing technology provides access to the pool of configurable resources including storage space, application, services and on demand network. Involvement of Cloud with the organization minimizes organization efforts toward fulfilling its customer’s needs. One of the major advantages of cloud computing is the Single Sign On (SSO) technique that allows the user to access multiple application services using a single user credential. In cloud computing, there are many issues, and challenges to be discussed. However, prevention from security attacks is much more difficult in preserving privacy of user agents. This paper has proposed SSO-based bio-metric authentication architecture for cloud computing services to overcome the security and privacy attacks. Bio-metric authentication is effective for resources controlled by end devices at the time of accessing the cloud services since these devices are computationally inefficient for user information processing during authentication. Accordingly, security attack in the cloud computing gets minimized using the proposed architecture. The proposed architecture also includes a novel approach in which there exist one to one relationship between user agent and the service provider. In this user agents can use their fingerprint while requesting for registration and accessing different cloud application services at cloud. Based on comparative study with several existing architectures, the highlights of the proposed architecture have been presented.

Semantic and Trade-Off Aware Location Privacy Protection in Road Networks Via Improved Multi-Objective Particle Swarm Optimization

Location privacy protection is an essential but challenging topic in the field of network security. Although the existing research methods, such as k-anonymity, mix zone, and differential privacy, show significant success, they usually neglect the location semantic and the proper trade-off between privacy and utility, which may allow attackers to obtain user privacy information by revealing the semantic correlation between the anonymous region and user's real location, thus causing privacy leakage. To solve this problem, we propose a location privacy protection scheme based on the k-anonymity technique, which provides practical location privacy-preserving through generating an anonymous set. This paper proposes a new location privacy attack strategy termed semantic relativity attack (SRA), which considers the location semantic problem. Correspondingly, a semantic and trade-off aware location privacy protection mechanism (STA-LPPM) is presented to achieve privacy protection with both high-level privacy and utility. To be specific, we model the location privacy protection as a multi-objective optimization problem and propose the Improved Multi-Objective Particle Swarm Optimization (IMOPSO) to generate the optimal anonymous set calculating the well-design fitness functions of the multi-objective optimization problem. In this way, the privacy scheme can provide mobile users with the right balance of privacy protection and service quality. Experiments reveal that our privacy scheme can effectively resist the semantic relativity attack while preventing significant utility degrading.

Applications of Differential Privacy in Social Network Analysis: A Survey

Differential privacy provides strong privacy preservation guarantee in information sharing. As social network analysis has been enjoying many applications, it opens a new arena for applications of differential privacy. This article presents a comprehensive survey connecting the basic principles of differential privacy and applications in social network analysis. We concisely review the foundations of differential privacy and the major variants. Then, we discuss how differential privacy is applied to social network analysis, including privacy attacks in social networks, models of differential privacy in social network analysis, and a series of popular tasks, such as analyzing degree distribution, counting subgraphs and assigning weights to edges. We also discuss a series of challenges for future work.

A Performance Study of Membership Inference Attacks on Different Machine Learning Algorithms

Nowadays, machine learning (ML) algorithms cannot succeed without the availability of an enormous amount of training data. The data could contain sensitive information, which needs to be protected. Membership inference attacks attempt to find out whether a target data point is used to train a certain ML model, which results in security and privacy implications. The leakage of membership information can vary from one machine-learning algorithm to another. In this paper, we conduct an empirical study to explore the performance of membership inference attacks against three different machine learning algorithms, namely, K-nearest neighbors, random forest, support vector machine, and logistic regression using three datasets. Our experiments revealed the best machine learning model that can be more immune to privacy attacks. Additionally, we examined the effects of such attacks when varying the dataset size. Based on our observations for the experimental results, we propose a defense mechanism that is less prone to privacy attacks and demonstrate its effectiveness through an empirical evaluation.

Pseudonym Mutable Based Privacy for 5G User Identity

Privacy, identity preserving and integrity have become key problems for telecommunication standards. Significant privacy threats are expected in 5G networks considering the large number of devices that will be deployed. As Internet of Things (IoT) and long-term evolution for machine type (LTE-m) are growing very fast with massive data traffic the risk of privacy attacks will be greatly increase. For all the above issues standards’ bodies should ensure users’ identity and privacy in order to gain the trust of service providers and industries. Against such threats, 5G specifications require a rigid and robust privacy procedure. Many research studies have addressed user privacy in 5G networks. This paper proposes a method to enhance user identity privacy in 5G systems through a scheme to protect the international mobile subscriber identity (IMSI) using a mutable mobile subscriber identity (MMSI) that changes randomly and avoids the exchange of IMSIs. It maintains authentication and key agreement (AKA) structure compatibility with previous mobile generations and improves user equipment (UE) synchronization with home networks. The proposed algorithm adds no computation overhead to UE or the network except a small amount in the home subscriber server (HSS). The proposed pseudonym mutable uses the XOR function to send the MMSI from the HSS to the UE which is reducing the encryption overhead significantly. The proposed solution was verified by ProVerif.

Significance of the Transition to Biometric Template Protection: Explore the Future

The IT security paradigm evolves from secret-based to biometric identity-based. Biometric identification has gradually become more popular in recent years for handheld devices. Privacy-preserving is a key concern when biometrics is used in authentication systems in the present world today. Nowadays, the declaration of biometric traits has been imposed not only by the government but also by many private entities. There are no proper mechanisms and assurance that biometric traits will be kept safe by such entities. The encryption of biometric traits to avoid privacy attacks is a giant problem. Hence, state-of-the-art safety and security technological solutions must be devised to prevent the loss and misuse of such biometric traits. In this paper, we have identified different cancelable biometrics methods with the possible attacks on the biometric traits and directions on possible countermeasures in order to design a secure and privacy-preserving biometric authentication system. We also proposed a highly secure method for cancelable biometrics using a non-invertible function based on Discrete Cosine Transformation and Index of max hashing. We tested and evaluated the proposed novel method on a standard dataset and achieved good results.

A Movie Recommendation System Based on Differential Privacy Protection

In the past decades, the ever-increasing popularity of the Internet has led to an explosive growth of information, which has consequently led to the emergence of recommendation systems. A series of cloud-based encryption measures have been adopted in the current recommendation systems to protect users’ privacy. However, there are still many other privacy attacks on the local devices. Therefore, this paper studies the encryption interference of applying a differential privacy protection scheme on the data in the user’s local devices under the assumption of an untrusted server. A dynamic privacy budget allocation method is proposed based on a localized differential privacy protection scheme while taking the specific application scene of movie recommendation into consideration. What is more, an improved user-based collaborative filtering algorithm, which adopts a matrix-based similarity calculation method instead of the traditional vector-based method when computing the user similarity, is proposed. Finally, it was proved by experimental results that the differential privacy-based movie recommendation system (DP-MRE) proposed in this paper could not only protect the privacy of users but also ensure the accuracy of recommendations.

Preserving Honest/Dishonest Users’ Operational Privacy with Blind Interference Calculation in Spectrum Sharing System

Spectrum sharing has been gaining its popular adoption as a potential solution to improve spectrum utilization in future wireless systems. Both Federal Communications Commission (FCC) and European Telecommunications Standards Institute (ETSI) support dynamic spectrum access (DSA) as an enabling technology for spectrum sharing. To effectively realize DSA in practice, users (from both defense and commercial sectors) are required to share their (radio) operational information, which risks exposing their security, privacy, and business plan to unintended agents. Protecting users' operating information is hence the key to DSA's success. In this paper, taking the FCC's spectrum access system (SAS) as a study case, we investigate the operational privacy issue of Incumbent Users (IUs) and honest/dishonest Secondary Users (SUs). For the case of IUs and honest SUs, we propose a privacy-preserving scheme for DSA by leveraging encryption and obfuscation methods (PSEO). To implement PSEO, we introduce an interference calculation scheme that allows users to calculate an interference budget without revealing operational information (e.g., antenna height, transmit power, location...), referred to as the blind interference calculation scheme (BICS). BICS also reduces the computing overhead of PSEO, compared with FCC's SAS by moving interference budgeting tasks to local users and calculating it in an offline manner. To further save the overhead in calculating the interference map, we introduce a quantization method and optimize the grid sizes of the terrestrial area of interest. Additionally, for the case of IUs and dishonest SUs, we propose a “punishment and forgiveness” (PF) mechanism, which draws support from SUs' reputation scores (RSs) and reputation histories (RHs), to encourage SUs to provide truthful information. Theoretical analysis and extensive simulations show that our proposed PSEO and PF-PSEO schemes can better protect all users' operational privacy under various privacy attacks, yielding higher spectrum utilization with less online overhead, compared with state of the art approaches.

Stay Connected, Leave no Trace

The intrinsic hardware imperfection of WiFi chipsets manifests itself in the transmitted signal, leading to a unique radiometric fingerprint. This fingerprint can be used as an additional means of authentication to enhance security. In fact, recent works propose practical fingerprinting solutions that can be readily implemented in commercial-off-the-shelf devices. In this paper, we prove analytically and experimentally that these solutions are highly vulnerable to impersonation attacks. We also demonstrate that such a unique device-based signature can be abused to violate privacy by tracking the user device, and, as of today, users do not have any means to prevent such privacy attacks other than turning off the device. We propose RF-Veil, a radiometric fingerprinting solution that not only is robust against impersonation attacks but also protects user privacy by obfuscating the radiometric fingerprint of the transmitter for non-legitimate receivers. Specifically, we introduce a randomized pattern of phase errors to the transmitted signal such that only the intended receiver can extract the original fingerprint of the transmitter. In a series of experiments and analyses, we expose the vulnerability of adopting naive randomization to statistical attacks and introduce countermeasures. Finally, we show the efficacy of RF-Veil experimentally in protecting user privacy and enhancing security. More importantly, our proposed solution allows communicating with other devices, which do not employ RF-Veil.