Privacy-Preserving Social Media Data Publishing for Personalized Ranking-Based Recommendation

Yang, Dingqi; Qu, Bingqing; Cudré-Mauroux, Philippe

doi:10.1109/tkde.2018.2840974

Cited by 72 publications

(33 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…9) represent whether a user's social status (defined in [45] as the number of followers divided by the number of followings) is greater than 5 or not. In [9], the social status is regarded as a private attribute that can be leaked by the user's behaviour and should be protected. We set 5 to be a threshold, because mobility patterns of people are substantially different when their social status is greater than 5 [45].…”

Section: A1 Experimental Comparison Among Attributes In Manhattanmentioning

confidence: 99%

Local Obfuscation Mechanisms for Hiding Probability Distributions

Kawamoto

Murakami²

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We introduce a formal model for the information leakage of probability distributions and define a notion called distribution privacy as the local differential privacy for probability distributions. Roughly, the distribution privacy of a local obfuscation mechanism means that the attacker cannot significantly gain any information on the distribution of the mechanism's input by observing its output. Then we show that existing local mechanisms can hide input distributions in terms of distribution privacy, while deteriorating the utility by adding too much noise. For example, we prove that the Laplace mechanism needs to add a large amount of noise proportionally to the infinite Wasserstein distance between the two distributions we want to make indistinguishable. To improve the tradeoff between distribution privacy and utility, we introduce a local obfuscation mechanism, called a tupling mechanism, that adds random dummy data to the output. Then we apply this mechanism to the protection of user attributes in location based services. By experiments, we demonstrate that the tupling mechanism outperforms popular local mechanisms in terms of attribute obfuscation and service quality.

show abstract

Section: A1 Experimental Comparison Among Attributes In Manhattanmentioning

confidence: 99%

Local Obfuscation Mechanisms for Hiding Probability Distributions

Kawamoto

Murakami²

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…The inference-prevention schemes are more focused on prevention by protectively processing and perturbing the location information prior to disclosure. The latter can be further broken down into location data perturbation techniques represented by [17,30,45,27,47,48] and trajectory inference prevention techniques represented by [7,8,34,35,36].…”

Section: Related Workmentioning

confidence: 99%

“…Location data perturbation schemes consists of perturbation through dummies [25,21], information-theoretic approaches [47,48], spatial location cloaking [5,11,17,19,24,30,45] and differential privacy [4,6,9,10,22,23]. The goal of location data perturbation is to perturb users' real location information so that the injected uncertainty can resist potential attacks made by adversaries.…”

Section: Related Workmentioning

confidence: 99%

“…The dummy trajectories should have similar properties of the real trajectories so that adversaries with map information cannot distinguish the real trajectories from the dummy trajectories. In information-theoretic approaches [47,48], to reduce privacy leakage, the location data of a user consists of public data that is safe to be released and private data that should be well protected. However, information about private data may be revealed from the public data.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Reversible spatio-temporal perturbation for protecting location privacy

Palanisamy

2019

Computer Communications

View full text Add to dashboard Cite

Ubiquitous deployment of low-cost mobile positioning devices and widespread use of high-speed wireless networks have resulted in a rapid growth of location-based applications. While location-based services provide a wide range of life enhancing experiences to users, the exposure of location information poses significant privacy risks that can invade the users' location privacy. Location privacy risks can be mitigated using location anonymization techniques that perturb the raw location of users to make the location indistinguishable from that of a set of other users. A fundamental limitation of traditional location anonymization techniques is that they are developed as unidirectional techniques that fail to support multi-level control of access to location data when data users have different access privileges on the exposed location information. As a result, location information once perturbed can not be reduced in terms of anonymity or degree of perturbation even when some data users have access to fine granular information in the exposed data. Recent techniques on reversible spatial cloaking techniques employ data anonymization keys to perturb a user's location in a pseudo-random manner such that the anonymized location information can be de-anonymized later using the anonymization keys. While reversible spatial cloaking provides support for multi-level location privacy, their performance is limited by their adopted spatial cloaking model in which the location perturbation occurs solely in the spatial domain without considering the temporal domain. Hence, reversible spatial location cloaking techniques obtain lower success rate and lower spatial resolution of the perturbed location leading to unreliable anonymization and lower service quality. This paper presents a new suite of reversible cloaking techniques that reversibly perturb location information of users using a spatio-temporal cloaking model, allowing data perturbation to occur along both spatial and temporal dimensions while still ensuring that the spatio-temporal expansion process is reversible when suitable access keys are provided. The proposed model achieves higher success rate and higher spatial resolution compared to reversible spatial cloaking. We compare our techniques through extensive experiments on real road networks. The results show that our techniques offer better QoS performance than the existing approaches and demonstrate strong attack resilience against adversarial attacks.

show abstract

“…(ii) when we know exact or approximate information on the input distributions (e.g., when we can use public datasets [11], [12] to learn approximate distributions of locations of male/female users if we want to obfuscate the attribute male/female). For the scenario (i), we clarify how much perturbation noise should be added to provide f -divergence DistP when we use an existing mechanism for obfuscating point data.…”

Section: Introductionmentioning

confidence: 99%

Local Distribution Obfuscation via Probability Coupling

Kawamoto

Murakami²

2019

2019 57th Annual Allerton Conference on Communication, Control, and Computing (Allerton)

View full text Add to dashboard Cite

We introduce a general model for the local obfuscation of probability distributions by probabilistic perturbation, e.g., by adding differentially private noise, and investigate its theoretical properties. Specifically, we relax a notion of distribution privacy (DistP) by generalizing it to divergence, and propose local obfuscation mechanisms that provide divergence distribution privacy. To provide f -divergence distribution privacy, we prove that probabilistic perturbation noise should be added proportionally to the Earth mover's distance between the probability distributions that we want to make indistinguishable. Furthermore, we introduce a local obfuscation mechanism, which we call a coupling mechanism, that provides divergence distribution privacy while optimizing the utility of obfuscated data by using exact/approximate auxiliary information on the input distributions we want to protect.in the scenario (i), an f -privacy mechanism can provide fdivergence DistP. Section VI generalizes DistP to use exact/approximate information on the input distribution in the scenario (ii), and proposes a local mechanism for providing DistP while optimizing utility. Section VII discusses related work and Section VIII concludes.

show abstract

Privacy-Preserving Social Media Data Publishing for Personalized Ranking-Based Recommendation

Cited by 72 publications

References 36 publications

Local Obfuscation Mechanisms for Hiding Probability Distributions

Local Obfuscation Mechanisms for Hiding Probability Distributions

Reversible spatio-temporal perturbation for protecting location privacy

Local Distribution Obfuscation via Probability Coupling

Contact Info

Product

Resources

About