At What Price? Exploring the Potential and Challenges of Differentially Private Machine Learning for Healthcare

2023 | HICSS | Citations: 1

Authors: Aslan, Aycan; Matschak, Tizian; Greve, Maike; Trang, Simon; Kolbe, Lutz

Abstract: The increased generation of data has become one of the main drivers of technolog ... Expand

Abstract: The increased generation of data has become one of the main drivers of technological innovation in healthcare. This applies in particular to the adoption of Machine Learning models that are used to generate value from the growing available healthcare data. However, the increased processing of sensitive healthcare data comes with challenges in terms of data privacy. Differential privacy, the method of adding randomness to the data to increase privacy, has gained popularity in the last few years as a possible solution. However, while the addition of randomness increases privacy, it also reduces overall model performance, generating a privacy-utility trade-off. Examining this trade-off, we contribute to the literature by providing an empirical paper that experimentally evaluates two prominent and innovative methods of differentially private Machine Learning on medical image and text data to deepen the understanding of the existing potential and challenges of such methods for the healthcare domain. Collapse

Topics: privacy differential privacy healthcare data knowledge base data protection

Methods: experiment computational algorithm machine learning blood pressure PyTorch

The Opportunity in Difficulty: A Dynamic Privacy Budget Allocation Mechanism for Privacy-Preserving Multi-dimensional Data Collection

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chen, Xue; Wang, Cheng; Yang, Qing; Hu, Teng; Jiang, Changjun

Abstract: Data collection under local differential privacy (LDP) has been gradually on the ... Expand

Abstract: Data collection under local differential privacy (LDP) has been gradually on the stage. Compared with the implementation of LDP on the single attribute data collection, that on multi-dimensional data faces great challenges as follows: (1) Communication cost. Multivariate data collection needs to retain the correlations between attributes, which means that more complex privatization mechanisms will result in more communication costs. (2) Noise scale. More attributes have to share the privacy budget limited by data utility and privacy-preserving level, which means that less privacy budget can be allocated to each of them, resulting in more noise added to the data. In this work, we innovatively reverse the complex multi-dimensional attributes, i.e., the major negative factor that leads to the above difficulties, to act as a beneficial factor to improve the efficiency of privacy budget allocation, so as to realize a multi-dimensional data collection under LDP with high comprehensive performance. Specifically, we first present a Multivariate k-ary Randomized Response (kRR) mechanism, called Multi-kRR. It applies the RR directly to each attribute to reduce the communication cost. To deal with the impact of a large amount of noise, we propose a Markov-based dynamic privacy budget allocation mechanism Markov-kRR, which determines the present privacy budget (flipping probability) of an attribute related to the state of the previous attributes. Then, we fix the threshold of flipping times in Markov-kRR and propose an improved mechanism called MarkFixed-kRR, which can obtain more optimized utility by choosing the suitable threshold. Finally, extensive experiments demonstrate the efficiency and effectiveness of our proposed methods. Collapse

Topics: privacy marketing management differential privacy data quality local differential privacy

Methods: computational algorithm experiment lasso regression archival research survey

Introduction to the Special Issue on Smart Systems for Industry 4.0 and IoT

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chen, Mu-Yen; Thuraisingham, Bhavani; Egrioglu, Erol; Rubio, Jose De Jesus

Abstract: INTRODUCTIONThe development of big data applications is driving the dramatic gr ... Expand

Abstract: INTRODUCTIONThe development of big data applications is driving the dramatic growth of hybrid data, often in the form of complex sets of cross-media content including text, images, videos, audios, and time series. Tremendous volumes of these heterogeneous data are derived from multiple IoT sources and present new challenges for the design, development, and implementation of effective information systems and decision support frameworks to meet heterogeneous computing requirements. Emerging technologies allow for the near real-time extraction and analysis of heterogeneous data to find meaningful information.Machine-learning algorithms allow computers to learn automatically, analyzing existing data to establish rules to predict outcomes of unknown data. However, traditional machine learning approaches do not meet the needs for Internet of Things (IoT) applications, calling for new technologies. Deep learning is a good example of emerging technologies that tackle the limitations of traditional machine learning through feature engineering, providing superior performance in highly complex applications. However, these technologies also raise new security and privacy concerns. Technology adoption and trust issues are of timely importance as well.Industrial operations are in the midst of rapid transformations, sometimes referred to as Industry 4.0, Industrial Internet of Things (IIoT), or smart manufacturing. These transformations are bringing fundamental changes to factories and workplaces, making them safer and more efficient, flexible, and environmentally friendly. Machines are evolving to have increased autonomy, and new human-machine interfaces such as smart tools, augmented reality, and touchless interfaces are making interaction more natural. Machines are also becoming increasingly interconnected within individual factories as well as to the outside world through cloud computing, enabling many opportunities for operational efficiency and flexibility in manufacturing and maintenance.An increasing number of countries have put forth national advanced manufacturing development strategies, such as Germany's Industry 4.0, the United States' Industrial Internet and manufacturing system based on CPS (Cyber-Physical Systems), and China's Internet Plus Manufacturing and Made in China 2025 initiatives. Smart Manufacturing aims to maximize transparency and access of all manufacturing process information across entire manufacturing supply chains and product lifecycles, with the Internet of Things (IoT) as a centerpiece to increase productivity and output value. This manufacturing revolution depends on technology connectivity and the contextualization of data, thus putting intelligent systems support and data science at the center of these developments. ACM Reference format: Collapse

Topics: production management computer data storage smart manufacturing differential privacy privacy

Methods: computational algorithm convolutional neural network machine learning

Reidentification Risk in Panel Data: Protecting for k-Anonymity

2023 | Information Systems Research | Citations: 0

Authors: Li, Shaobo; Schneider, Matthew J.; Yu, Yan; Gupta, Sachin

Abstract: We consider the risk of reidentification of panelists in marketing research data ... Expand

Abstract: We consider the risk of reidentification of panelists in marketing research data that are widely used to obtain insights into buyer behavior and to develop marketing strategy. We find that 17%–94% of the panelists in 15 frequently bought consumer goods categories are subject to high risk of reidentification through a potential record linkage attack based on their unique purchasing histories even when their identities are anonymized. We first demonstrate that the risk of reidentification is vastly understated by unicity, the conventional measure. Instead, we propose a new measure of reidentification risk, termed sno-unicity, which accounts for the longitudinal nature of panel data, and show that it is much larger than unicity. To protect the privacy of panelists, we consider the well-known privacy notion of k-anonymity and develop a new approach called graph-based minimum movement k-anonymization (k-MM) that is designed especially for panel data. The proposed k-MM approach can be formulated as an optimization problem in which the objective is to minimally distort variables in the original data based on weights that users prespecify corresponding to their use case. We further show how our approach can be extended to achieve l-diversity. We apply the k-MM approach to two different panel data sets that are widely used in marketing research. To achieve a given privacy level, compared with several benchmark protection methods, the protected data from our method result in the least distortion in inferences about key marketing metrics, such as brand market shares, share of category requirements, brand switching rates, and marketing-mix parameters estimated from a hierarchical Bayesian brand choice model. History: Param Singh, Senior Editor and Associate Editor. Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2022.1169. Collapse

Topics: privacy marketing management differential privacy data protection record linkage

Methods: longitudinal research optimization model descriptive statistic simulation cluster analysis

Unleashing The Potential of Data Ecosystems: Establishing Digital Trust through Trust-Enhancing Technologies

2023 | European Conference On Information Systems | Citations: 0

Authors: Schäfer, Fabian; Rosen, Jeremy; Zimmermann, Christian; Wortmann, Felix

Abstract: Companies increasingly innovate data-driven business models, enabling them to c ... Expand

Abstract: Companies increasingly innovate data-driven business models, enabling them to create new products and services. Emerging data ecosystems provide these companies access to complementary data, offering them additional potential. This, however remains untapped, as a lack of digital trust prevents companies from sharing data within these ecosystems. By using trust-enhancing technologies, companies can establish trust; this can be explained through the theoretical lens of system trust. Using a design science research approach helped us to unlock the knowledge of 21 experts and identify five technologies with the potential to solve the trust challenge: self-sovereign identities, differential privacy, fully homomorphic encryption, trusted execution environments and secure multiparty computation. We integrated these technologies into the data sharing process in data ecosystems and elaborated on their limitations and maturity. Ultimately, we derived two principles that allow for adapting our results to future technological developments: complementarity and customization. Collapse

Topics: digital ecosystem differential privacy fully homomorphic encryption distributed ledger technology data encryption

Methods: design science qualitative interview design artifact design requirement design process

What are the trend and core knowledge of information security? A citation and co-citation analysis

2023 | Information & Management | Citations: 0

Authors: Shiau, Wen-Lung; Wang, Xiaoqun; Zheng, Fei

Abstract: Information technology brings business success opportunities, but also causes po ... Expand

Abstract: Information technology brings business success opportunities, but also causes potential safety hazards to orga­ nizations. In response to the increasing academia and industry concerns regarding information security (ISec), this study systematically explored extant ISec research and identified eight core knowledge groups, including (1) intrusion detection, (2) privacy protection, (3) secure machine learning, (4) cryptosystem, (5) data service se­ curity, (6) malware analysis, (7) security decision-making, and (8) security management. The detection of research hotspots shows that data service security and risk management garner the most current research attention. Furthermore, we establish a comprehensive ISec framework to help systematically understand and achieve ISec. Collapse

Topics: privacy decision making intrusion detection system malware database system

Methods: co-citation analysis cluster analysis literature sample hierarchical clustering multidimensional scaling

MEDIATING THE TENSION BETWEEN DATA SHARING AND PRIVACY: THE CASE OF DMA AND GDPR

2023 | European Conference On Information Systems | Citations: 0

Authors: Weigl, Linda; Barbereau, Tom; Sedlmeir, Johannes; Zavolokina, Liudmila

Abstract: The Digital Markets Act (DMA) constitutes a crucial part of the European legisl ... Expand

Abstract: The Digital Markets Act (DMA) constitutes a crucial part of the European legislative framework addressing the dominance of 'Big Tech'. It intends to foster fairness and competition in Europe's digital platform economy by imposing obligations on 'gatekeepers' to share end-user-related information with business users. Yet, this may involve the processing of personal data subject to the General Data Protection Regulation (GDPR). The obligation to provide access to personal data in a GDPR-compliant manner poses a regulatory and technical challenge and can serve as a justification for gatekeepers to refrain from data sharing. In this research-in-progress paper, we analyze key tensions between the DMA and the GDPR through the paradox perspective. We argue through a task-technology fit approach how privacyenhancing technologies -particularly anonymization techniques -and portability could help mediate tensions between data sharing and privacy. Our contribution provides theoretical and practical insights to facilitate legal compliance. Collapse

Topics: legal context General Data Protection Regulation privacy differential privacy data protection

Methods: decision tree classification structured literature research computational algorithm synthesis qualitative interview

Theories: theory of task technology fit

Play Your Cards Right: Utilizing Differential Privacy as a Competitive Advantage

2022 | Americas Conference on Information Systems | Citations: 0

Authors: Diesterhöft, Till Ole; Aslan, Aycan; Braun, Marvin

Abstract: Given the increasing amount of information being leveraged in business models, ... Expand

Abstract: Given the increasing amount of information being leveraged in business models, a growing number of companies have adopted differential privacy (DP) as a means of ensuring privacy. DP enables the adjustment of privacy characteristics of individual information. Nevertheless, measurements taken rarely lead to total prevention of information loss due to external events, i.e., data breaches. Owing to the control nature of DP, we argue that even in the occurrence of a data breach, investments made in this technology can be strategically exploited to positively influence perceived information privacy control. Drawing on prior literature, we propose a relationship between customers' information privacy control and satisfaction. By conducting a between-subjects experiment (n=185), we can confirm these assumptions. Our paper paves the way for the disclosure of privacy measures, in particular DP, to enable a more positive reaction of customers towards data breaches, thus creating competitive advantages. KeywordsDifferential privacy, perceived information control, data breach notifications. Collapse

Topics: data breach privacy mobile application information privacy concern customer satisfaction

Methods: partial least squares path modeling experiment survey statistical hypothesis test PLS tool

Privacy and Confidentiality in Process Mining: Threats and Research Challenges

2022 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Elkoumy, Gamal; Fahrenkrog-Petersen, Stephan A.; Sani, Mohammadreza Fani; Koschmider, Agnes; Mannhardt, Felix; Von Voigt, Saskia Nuñez; Rafiei, Majid; Waldthausen, Leopold Von

Abstract: Privacy and confidentiality are very important prerequisites for applying proces ... Expand

Abstract: Privacy and confidentiality are very important prerequisites for applying process mining to comply with regulations and keep company secrets. This article provides a foundation for future research on privacy-preserving and confidential process mining techniques. Main threats are identified and related to a motivation application scenario in a hospital context as well as to the current body of work on privacy and confidentiality in process mining. A newly developed conceptual model structures the discussion that existing techniques leave room for improvement. This results in a number of important research challenges that should be addressed by future process mining research. Collapse

Topics: privacy process mining General Data Protection Regulation differential privacy legal context

Methods: event log business process modeling conceptual modelling literature study computational algorithm

Understanding Cyberprivacy: Context, Concept, and Issues

2022 | International Conference on Business Informatics | Citations: 0

Authors: Eltahawy, Bahaa; Dang, Duong

Abstract: Cyberprivacy has become one of the most worrisome issues in the age of digitaliz ... Expand

Abstract: Cyberprivacy has become one of the most worrisome issues in the age of digitalization, as data breaches have increased at an alarming rate, and the development of technology has changed privacy norms themselves. Thus, maintaining cyberprivacy is important for both academia and practitioners. However, the literature on cyberprivacy is fragmented, since the topic is multidisciplinary and often confused with cybersecurity and data privacy. In this study, we seek to understand cyberprivacy by conducting a comprehensive literature review and analyzing 79 selected articles on the topic between 2008 and 2021. Our analysis shows that there are eight contexts associated with cyberprivacy. We proposed concepts on cyberprivacy from different views and highlighted four issues related to cyberprivacy for future consideration. Taken together, the knowledge on cyberprivacy, its challenges and its practices does not seem to accumulate. Consequently, there is a need for more targeted research on the topic to cover different contexts. Collapse

Topics: privacy legal context differential privacy database system blockchain

Methods: literature sample literature study