Privacy Assurance Research Articles

Federated Learning for Clients’ Data Privacy Assurance in Food Service Industry

The food service industry must ensure that service facilities are free of foodborne pathogens hosted by organic residues and biofilms. Foodborne diseases put customers at risk and compromise the reputations of service providers. Fluorescence imaging, empowered by state-of-the-art artificial intelligence (AI) algorithms, can detect invisible residues. However, using AI requires large datasets that are most effective when collected from actual users, raising concerns about data privacy and possible leakage of sensitive information. In this study, we employed a decentralized privacy-preserving technology to address client data privacy issues. When federated learning (FL) is used, there is no need for data sharing across clients or data centralization on a server. We used FL and a new fluorescence imaging technology and applied two deep learning models, MobileNetv3 and DeepLabv3+, to identify and segment invisible residues on food preparation equipment and surfaces. We used FedML as our FL framework and Fedavg as the aggregation algorithm. The model achieved training and testing accuracies of 95.83% and 94.94% for classification between clean and contamination frames, respectively, and resulted in intersection over union (IoU) scores of 91.23% and 89.45% for training and testing, respectively, of segmentation of the contaminated areas. The results demonstrated that using federated learning combined with fluorescence imaging and deep learning algorithms can improve the performance of cleanliness auditing systems while assuring client data privacy.

A Federated Learning System for Histopathology Image Analysis with an Orchestral Stain-Normalization GAN.

Currently, data-driven based machine learning is considered one of the best choices in clinical pathology analysis, and its success is subject to the sufficiency of digitized slides, particularly those with deep annotations. Although centralized training on a large data set may be more reliable and more generalized, the slides to the examination are more often than not collected from many distributed medical institutes. This brings its own challenges, and the most important is the assurance of privacy and security of incoming data samples. In the discipline of histopathology image, the universal stain-variation issue adds to the difficulty of an automatic system as different clinical institutions provide distinct stain styles. To address these two important challenges in AI-based histopathology diagnoses, this work proposes a novel conditional Generative Adversarial Network (GAN) with one orchestration generator and multiple distributed discriminators, to cope with multiple-client based stain-style normalization. Implemented within a Federated Learning (FL) paradigm, this framework well preserves data privacy and security. Additionally, the training consistency and stability of the distributed system are further enhanced by a novel temporal self-distillation regularization scheme. Empirically, on large cohorts of histopathology datasets as a benchmark, the proposed model matches the performance of conventional centralized learning very closely. It also outperforms state-of-the-art stain-style transfer methods on the downstream Federated Learning image classification task, with an accuracy increase of over 20.0% in comparison to the baseline classification model.

DPrio: Efficient Differential Privacy with High Utility for Prio

Private data collection systems such as Prio ensure data privacy by distributing trust among a set of mutually trusted parties, to allow for aggregate data collection without disclosing any single client's data in the clear. While systems like Prio are undergoing widespread interest and adoption, these systems lack efficient mechanisms to provide differential privacy guarantees. In this work, we present a lightweight method that we call DPrio to augment Prio and related systems with differential privacy assurances while ensuring higher data utility than existing noise generation protocols. We compare our results against four related constructions in the literature, and identify how DPrio achieves improved data utility relative to the assumed number of dishonest clients and servers, with only minimal (and batchable) server communication overhead. We present several case studies and discuss considerations for real-world implementations.

Tool-supported method for privacy analysis of a business process model

The shift to data-driven decision-making brings opportunities to organisations. However, reliance on sensitive data about identifiable persons poses the obligation to cope with data privacy management with respect to the local legislation. As the privacy protection regulation (i.e. GDPR) in the EU has gained its power relatively recently, no established procedures or frameworks guide privacy analysis and assurance. The paper presents a tool-supported method for privacy analysis of a business process model. The proposed method aims to support the elicitation of requirements to the information system to comply with GDPR. Additionally, the method supports the selection of technical measures for privacy assurance based on their effectiveness in the context of a business process. The method’s usability is validated by the experimental application of the method to a ride fulfilment process in a ride-hailing company enabled by an autonomous driving system.

Privacy-Preserving Industrial Control System Anomaly Detection Platform

With the development of IT technologies, an increasing number of industrial control systems (ICSs) can be accessed from the public Internet (with authentication). In such an open environment, cyberattacks become a serious threat to both ICS system integrity and data privacy. As a countermeasure, anomaly detection systems are often deployed to analyze the network traffic. However, due to privacy regulation, the network packages cannot be directly processed in plaintext in many countries. In this work, we present a privacy-preserving anomaly detection platform for ICS. The platform consists of three nodes running low-latency MPC protocols to evaluate the live network packages using decision trees on the fly with privacy assurance. Our benchmark result shows that the platform can process thousands of packages every ten seconds.

Techniques and protocols for enhancing data privacy in cloud computing: A review

The emergence of cloud computing has enabled millions of devices across the globe to communicate and exchange massive amounts of data that is classified as private with each other. Given that these devices and platforms are scattered in different geographical areas, communicating with IoTs all the time, the issue of data security, in particular, assurance of data privacy has arisen. Several attempts in mitigating this issue has been by both academic and industry researchers in terms of technology, practices and protocols but it still remains a serious challenge even as cloud computing forges into the future. As a result, both individuals and organizations whose private data are held in the cloud are getting jittery due to the frequent data compromise or exposure in the cloud servers. This paper reviews the various challenges in data privacy and the mechanisms and techniques proposed by various researchers in bid to mitigate them and create a trustworthy cloud environment. It undertakes a comparative research analysis of the existing research work on data privacy protection techniques that have been proposed and implemented in the cloud computing environment.

Small scale IoT device privacy evaluation using Petri net modeling

Most Small Scale IoT (SSIoT) devices on the market gather a significant amount of sensitive information, yet many lack privacy controls, introducing significant privacy and safety risk to users. Such risks stem from the lack of privacy integration into the system development process. No formalized SSIoT data flow model currently integrates privacy elements for evaluation during the system development lifecycle (SDLC). This work aims to review current data flow modeling techniques, used in most SSIoT System Development Lifecycle (SDLC), to identify privacy gaps and assess requisite privacy controls necessary to improve user privacy. To verify this, we designed a simulation experiment using Petri net to evaluate the current privacy controls and hotspots during SSIoT data transitions. We assess our Petri net model using a Barbie Smart connected toy user interaction. The results show that Petri net has unique privacy elements and verification schemes over all other data flow modeling techniques. Further, it provides privacy assurance, evaluates privacy by identifying privacy hotspots needing controls, and minimizes privacy-related risks such as breach of personally identifiable information and interaction data during SSIoT device use.

DETECTING SUPPLIER PAYMENT ERRORS AND FRAUD USING A DATA WAREHOUSE AUDIT APPROACH

ABSTRACT This study provides management, accountants, and auditors with immediately applicable knowledge on supplier payment error discovery using a data warehouse. Despite the universality of supplier payment processing across government and non-government sectors, the associated error risks have received little attention in the academic literature and only limited attention in the practitioner literature. This paper seeks to articulate the data mining methods needed to perform such audits. A case study approach was used to analyze 15 client auditees in five countries using a variety of accounts payable systems. Action research methodology brought rigor to the development of incremental knowledge of error hypotheses and detection techniques. In this way, the prior knowledge gained from each case enriched the subsequent iterations. Even among clients whose management presumed human and system controls were robust, findings showed an average 0.19 percent error rate in payments to suppliers. These were mostly recoverable as cash refunds or credit notes. The audit cost-to-benefit ratio averaged 1:33, suggesting clients received financial returns 33 times higher than the cost of the audit, on average. Concurrently, internal controls were also hardened in accounts payable, tax compliance, and fraud prevention to secure additional future cost savings. Besides material cash recoveries and the occasional discovery of fraud, incremental lessons were also learned in data warehouse construction, data cleansing, privacy assurance, and tax compliance.

Blockchain for Precision Medicine

Because evidence generation for precision medicine involves highly complex bioinformatics on human genomic and other molecular data sets, the reproducibility of the evidence generation and the sharing of highly sensitive raw genomic data present major challenges for life sciences sponsors and the FDA. Blockchain offers solutions for reproducibility and privacy assurance in precision medicine.

Internet of Things (Iot) and Cyber-Physical Systems for Distributed Sensing and Control

The combination of Cyber-Physical Systems (CPS) and the Internet of Things (IoT) has resulted in a massive increase in the amount of real-time data that is easily available. As a result, this combination has made it possible to obtain previously inaccessible insights into a variety of systems and processes. This has helped us make better judgements, increase productivity, and improve safety and security across a wide range of industries, from manufacturing to medical. Increasing productivity and efficiency is one of the biggest advantages offered by IoT and CPS. Data collection on various systems and processes might reveal inefficiencies and areas for improvement. We can anticipate savings and higher productivity as a consequence. As an extra benefit, real-time data monitoring and analysis enables us to identify any risks to public health and safety and take appropriate action. As a result, accidents can be avoided. However, the widespread adoption of IoT and CPS presents a number of difficulties, including the management of enormous amounts of data and the assurance of privacy and security. Despite these drawbacks, the potential benefits of these technologies are currently too great to ignore. If we continue to support R&D, we may make significant advancements in industries like industrial automation, precision agriculture, and healthcare. It will be vital to overcome these problems if IoT and CPS are to realise their full potential. These challenges include the requirement for better security measures and the creation of increasingly complex data processing algorithms. We can contribute to a safer, wealthier, and more connected society by acting in this manner.

IPrivJoin: An ID-Private Data Join Framework for Privacy-Preserving Machine Learning

The world has observed an increasing trend in the development of Privacy-Preserving Machine Learning (PPML) for cross-silo collaborative model training over sensitive data. As the first essential step of cross-silo PPML, it is critical that the parties can align their dataset with privacy assurance, i.e., <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">private data join</i> . However, the existing private data join methods typically leak the ID information in the dataset intersection, which often raises privacy concerns. In this work, we propose iPrivJoin: a novel framework of ID-private data join for PPML. Compared with naively using circuit-based Private Set Intersection (circuit-PSI) for data join, the proposed framework has two advantages. (i) data volume reduction. iPrivJoin utilizes oblivious shuffle to securely trim off the redundant data that is outside the intersection, while the entire dataset needs to be carried to further process in the circuit-PSI based approach. (ii) efficiency improvement. iPrivJoin introduces a new private encoding technique to avoid the expensive circuit evaluation that is needed in circuit-PSI. As a result, compared with directly using circuit-PSI, PPML with iPrivJoin enjoys approximately 3× of speedup. Moreover, we propose a new oblivious shuffle protocol, which may be of independent interest. It achieves 1.44× of speedup to the state-of-the-art in the real-world WAN network setting.

Privacy-Preserving Statistical Analysis of Genomic Data Using Compressive Mechanism with Haar Wavelet Transform.

To promote the use of personal genome information in medicine, it is important to analyze the relationship between diseases and the human genomes. Therefore, statistical analysis using genomic data is often conducted, but there is a privacy concern with respect to releasing the statistics as they are. Existing methods to address this problem using the concept of differential privacy cannot provide accurate outputs under strong privacy guarantees, making them less practical. In this study, for the first time, we investigate the application of a compressive mechanism to genomic statistical data and propose two approaches. The first is to apply the normal compressive mechanism to the statistics vector along with an algorithm to determine the number of nonzero entries in a sparse representation. The second is to alter the mechanism based on the data, aiming to release significant single nucleotide polymorphisms with a high probability. In this algorithm, we apply the compressive mechanism with the input as a sparse vector for significant data and the Laplace mechanism for nonsignificant data. By using the Haar wavelet transform for the compressive mechanism, we can determine the number of nonzero elements and the amount of noise. In addition, we give theoretical guarantees that our proposed methods achieve ϵ-differential privacy. We evaluated our methods in terms of accuracy and rank error compared with the Laplace and exponential mechanisms. The results show that our second method in particular can guarantee high privacy assurance as well as utility.

Do privacy assurances work? a study of truthfulness in healthcare history data collection.

Patients often provide untruthful information about their health to avoid embarrassment, evade treatment, or prevent financial loss. Privacy disclosures (e.g. HIPAA) intended to dissuade privacy concerns may actually increase patient lying. We used new mouse tracking-based technology to detect lies through mouse movement (distance and time to response) and patient answer adjustment in an online controlled study of 611 potential patients, randomly assigned to one of six treatments. Treatments differed in the notices patients received before health information was requested, including notices about privacy, benefits of truthful disclosure, and risks of inaccurate disclosure. Increased time or distance of device mouse movement and greater adjustment of answers indicate less truthfulness. Mouse tracking revealed a significant overall effect (p&lt;0.001) by treatment on the time to reach their final choice. The control took the least time indicating greater truthfulness and the privacy + risk group took the longest indicating least truthfulness. Privacy, risk, and benefit disclosure statements led to greater lying. These differences were moderated by gender. Mouse tracking results largely confirmed the answer adjustment lie detection method with an overall treatment effect (p &lt; .0001) and gender differences (p &lt; .0001) on truthfulness. Privacy notices led to decreased patient honesty. Privacy notices should perhaps be administered well before personal health disclosure is requested to minimize patient untruthfulness. Mouse tracking and answer adjustment appear to be health care lie-detection methods to enhance optimal diagnosis and treatment.

Behavioral pre-testing of COVID Tracker, Ireland’s contact-tracing app.

Contact-tracing mobile phone apps can play a role in controlling the spread of COVID-19, but their success hinges on widespread public acceptance, uptake and use, which are difficult for public administrators to foresee. We report on a rapid behavioral pre-test of COVID Tracker, Ireland’s contact-tracing app, prior to its launch. A large sample of participants were randomized to receive different versions of a trial app. They responded to an online survey while downloading and using the app on their phones in real time. Experimental manipulations focused on: (i) the level of privacy assurance provided in the app, (ii) the goal-framing of the purpose of the app and (iii) the structuring of the exposure notification. Almost one in five participants mentioned privacy concerns in relation to their likelihood of downloading the app, but concerns were lower and engagement stronger in a condition that included additional assurances regarding the privacy of users’ data in the app. This finding informed the final version of the app. Overall, our results demonstrate the value and feasibility of pre-testing digital interfaces to improve public administration.

Privacy assurances and social sharing in social commerce: The mediating role of threat-coping appraisals

The social sharing of consumers (e.g., posting about a shopping experience) is the basis for the flourishing of social commerce; however, it conflicts with consumers’ psychological need for privacy. This study attempts to investigate the comprehensive effect of privacy assurances from different institutions on social sharing, involving privacy policy and privacy-protective technology provided by social media companies as well as government regulation and industry self-regulation. We conducted a large-scale empirical survey with 1,311 users of Xiaohongshu in China. The results show that all the four privacy assurances have a positive effect on social sharing that is almost equal. The results also reveal that threat appraisal and coping appraisal jointly mediate the relationship of privacy assurances and social sharing. Our results provide operable implications for stakeholders on how to facilitate social sharing and thus the flourishing of s-commerce through fostering privacy assurances.

Privacy risk assessment and privacy-preserving data monitoring

Privacy regulations press organisations to handle personal data with reinforced caution. Moreover, organisations are dealing with increasing amounts of Personally Identifiable Information in their systems. Thus, there is a high demand not only for privacy-preserving data processing mechanisms but also privacy-enhancing services. As such, we propose the Personal Data Analyser, a tool that increases privacy assurances and minimises privacy risks through automated privacy-preserving data monitoring and privacy risk assessment mechanisms. Automated data monitoring is achieved with a hybrid mechanism that employs Regular Expressions, Natural Language Processing tools, and machine learning models such as Multilayer Perceptron and Random Forests. Our privacy risk assessment mechanism is based on custom-built crisp and fuzzy models, that consider information such as data processor reputation, data sensitiveness and other inputs in order to assess privacy risk associated with data transactions. Our work is integrated and validated under real use cases of the PoSeID-on platform and warns users whenever potential privacy risks are detected. Validation under PoSeID-on’s pilots and its users proved beneficial not only to assess our solution but also to raise users’ awareness of their data. The results of this work show that our solution is an effective Privacy Enhancing Technology that increases privacy assurances between organisations and their users.

Hotel privacy management and guest trust building: A relational signaling perspective

By using the tenets of relational signaling theory, this study proposes a conceptual model that investigates the effect of hotel privacy management (thin trust) on guests’ cognitive and affective trust (thick trust), which influences guest intention to accept personal information collection and switch to another hotel. Survey data was collected online and analyzed by using confirmatory factory analysis and structural equation modeling. The results reveal that hotels’ privacy policy, privacy assurance, and employees' information access control are the key factors forming guests’ cognitive and affective trust. While both cognitive and affective trust increase guests’ intention to accept information collection by hotels, only cognitive trust significantly decreases guests’ switching intention. This study also reveals that when hotels hold weaker reputation both cognitive and affective trust play a critical role in increasing guests’ acceptance of hotel information collection. Theoretical implications to the privacy-related literature and practical implications to hotel practitioners are discussed.

Achieving Efficient and Privacy-Preserving Cross-Domain Big Data Deduplication in Cloud

Secure data deduplication can significantly reduce the communication and storage overheads in cloud storage services, and has potential applications in our big data-driven society. Existing data deduplication schemes are generally designed to either resist brute-force attacks or ensure the efficiency and data availability, but not both conditions. We are also not aware of any existing scheme that achieves accountability, in the sense of reducing duplicate information disclosure (e.g., to determine whether plaintexts of two encrypted messages are identical). In this paper, we investigate a three-tier cross-domain architecture, and propose an efficient and privacy-preserving big data deduplication in cloud storage (hereafter referred to as EPCDD). EPCDD achieves both privacy-preserving and data availability, and resists brute-force attacks. In addition, we take accountability into consideration to offer better privacy assurances than existing schemes. We then demonstrate that EPCDD outperforms existing competing schemes, in terms of computation, communication and storage overheads. In addition, the time complexity of duplicate search in EPCDD is logarithmic.

Mobile app users' privacy concerns: different heuristics for privacy assurance statements in the EU and China

PurposeAs mobile apps request permissions from users, protecting mobile users' personal information from being unnecessarily collected and misused becomes critical. Privacy regulations, such as General Data Protection Regulation in the European Union (EU), aim to protect users' online information privacy. However, one’s understanding of whether these regulations effectively make mobile users less concerned about their privacy is still limited. This work aims to study mobile users' privacy concerns towards mobile apps by examining the effects of general and specific privacy assurance statements in China and the EU.Design/methodology/approachDrawing on ecological rationality and heuristics theory, an online experiment and a follow-up validation experiment were conducted in the EU and China to examine the effects of privacy assurance statements on mobile users' privacy concerns.FindingsWhen privacy regulation is presented, the privacy concerns of Chinese mobile users are significantly lowered compared with EU mobile users. This indicates that individuals in the two regions react differently to privacy assurances. However, when a general regulation statement is used, no effect is observed. EU and Chinese respondents remain unaffected by general assurance statements.Originality/valueThis study incorporates notions from fast and frugal heuristics end ecological rationality – where seemingly irrational decisions may make sense in different societal contexts.

Exploring the fit between mobile application service and application privacy

PurposeWith the popularity of mobile applications and increasing consumer awareness of application privacy, this paper aims to introduce a new construct of service-privacy fit (i.e. the perceived degree of match between the service of a mobile application and a privacy permission request) to predict consumers’ mobile application adoption.Design/methodology/approachFour experiments were carried out to test the hypothesized relationships. The first study investigated the direct impact of service-privacy fit on application adoption and the mediators of benefit expectancy and privacy concerns. The second, third and fourth studies further applied the moderated mediation model to analyze the moderating roles of benefit message type, regulatory focus type and privacy assurance.FindingsThe results show that service-privacy fit influences application adoption not only directly but also indirectly via the mediators of benefit expectancy and privacy concerns. Furthermore, the findings confirm the moderators of benefit message type, regulatory focus type and privacy assurance.Originality/valueDrawn from the perspectives of task-technology fit and information boundary theory, this paper introduces a new construct of service-privacy fit as a determinant of application adoption. Grounded in privacy calculus theory, it further explains this relationship through mediating effects of benefit expectancy and privacy concerns. Furthermore, this paper proposes that benefit messages and privacy assurance are effective coping strategies to increase the benefit expectancy and reduce the privacy concerns of applications. Based on the perspective of regulatory fit theory, this study further shows that the effects of coping strategies rely on personal traits. The findings enrich the existing knowledge of mobile application adoption and application privacy, suggesting that practitioners should consider mobile consumers’ perception of service-privacy fit when developing applications.