VACCINE: Using Contextual Integrity For Data Leakage Detection

Shvartzshnaider, Yan; Pavlinovic, Zvonimir; Balashankar, Ananth; Wies, Thomas; Subramanian, Lakshminarayanan; Nissenbaum, Helen; Mittal, Prateek

doi:10.1145/3308558.3313655

Cited by 22 publications

(9 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides this diversity in communication styles, Enron's code of communication is publicly available and allows us to easily draw contextual integrity norms. This approach has also been used previously to assess the performance of data leakage prevention systems [46]. However, using such a dataset for a privacy preservation system may raise a challenging ethical issue.…”

Section: Discussion and Limitationsmentioning

confidence: 99%

“…Although this corpus was originally leaked in the scope of a financial scandal, it is nowadays a publicly available dataset spreading over several years that contains email conversations at individual, group, and company level. As such, it has been used in prior literature to evaluate the efficiency of privacy-leakage prevention systems [46]. Through the offline study of this corpus, we aim to answer the following questions:…”

Section: Localmentioning

confidence: 99%

“…(3) How effective is Aquilis' Cross Platform Module to measure privacy loss incurred due to use sharing data through multiple platforms? We follow the protocols adopted by Shvartzshnaider et al [46] to select email conversations for our experiment. An email conversation is defined as a set of emails with the same subject (modulo reply and forwarding prefixes).…”

Section: Localmentioning

confidence: 99%

“…Masood et al [27] propose Incognito, a method for Web data risk prediction combined with an obfuscation technique to protect the users' privacy by adding noise to the data. VACCINE [46] proposes a design methodology for data leakage prevention based on the theory of contextual integrity (CI). However, these methods primarily target Desktop machines.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Aquilis

Kumar

Braud

Kwon

et al. 2020

Proc. ACM Interact. Mob. Wearable Ubiquitous Technol.

View full text Add to dashboard Cite

Smartphones are nowadays the dominant end-user device. As a result, they have become gateways to all users' communications, including sensitive personal data. In this paper, we present Aquilis, a privacy-preserving system for mobile platforms following the principles of contextual integrity to define the appropriateness of an information flow. Aquilis takes the form of a keyboard that reminds users of potential privacy leakages through a simple three-colour code. Aquilis considers the instantaneous privacy risk related to posting information (Local Sensitivity), the risk induced by repeating information over time (Longitudinal Sensitivity) and on different platforms (Cross-platform Sensitivity). Considering 50% of Aquilis warnings decreases the proportion of inappropriate information by up to 30%. Repeating information over time or in a broader exposure context increases the risk by 340% in a one-to-one context. We develop our own labeled privacy dataset of over 1000 input texts to evaluate Aquilis' accuracy. Aquilis significantly outperforms other state-of-the-art methods (F-1-0.76). Finally, we perform a user study with 35 highly privacy-aware participants. Aquilis privacy metric is close to users' privacy preferences (average divergence of 1.28/5). Users found Aquilis useful (4.41/5), easy to use (4.4/5), and agreed that Aquilis improves their online privacy awareness (4.04/5).

show abstract

Section: Discussion and Limitationsmentioning

confidence: 99%

Section: Localmentioning

confidence: 99%

Section: Localmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Aquilis

Kumar

Braud

Kwon

et al. 2020

Proc. ACM Interact. Mob. Wearable Ubiquitous Technol.

View full text Add to dashboard Cite

show abstract

“…Today, the organizations are being harmed by data that is exposed to parties that are unauthorised [1]. These data leaks may result in harm in different ways.…”

Section: Introductionmentioning

confidence: 99%

Data Leak Identification in Social Networks using K Means Clustering & Tabu K Means Clustering

Karthik¹,

Tamizhazhagan²,

Narayana³

2019

IJITEE

View full text Add to dashboard Cite

The prevention of leakage of data has been defined as a process or solution which identifies data that is confidential, tracks the data in a way in which it moves in and out of its enterprise to prevent any unauthorized data disclosure in an intentional or an unintentional manner. As data that is confidential is able to reside on various computing devices and move through several network access points or different types of social networks such as emails. Leakage of emails has been defined as if the email either deliberately or accidentally goes to an addressee to whom it should not be addressed. Data Leak Prevention (DLP) is the technique or product that tries mitigating threats to data leaks. In this work, the technique of clustering will be combined with the frequency of the term or the inverse document frequency in order to identify the right centroids for analysing the various emails that are communicated among members of an organization. Every member will fit in to various topic clusters and one such topic cluster can also comprise of several members in the organization who have not communicated with each other earlier. At the time when a new email is composed, every addressee will be categorized to be a potential leak recipient or one that is legal. Such classification was based on the emails sent among the sender and the receiver and also on their topic clusters. The work had investigated the technique of K-Means clustering and also proposed a Tabu - K-Means (TABU-KM) technique of clustering to identify points of optimal clustering. The proposed TABU-KM optimizes the K-Means clustering. Experimental results demonstrated that the proposed method achieves higher True Positive Rate (TPR) for known and unknown recipient and lower False Positive Rate (FPR) for known and unknown recipient

show abstract

Depthwise Separable Convolutional Neural Network for Confidential Information Analysis

Jiang

et al. 2020

Knowledge Science, Engineering and Management

View full text Add to dashboard Cite

VACCINE: Using Contextual Integrity For Data Leakage Detection

Cited by 22 publications

References 24 publications

Aquilis

Aquilis

Data Leak Identification in Social Networks using K Means Clustering & Tabu K Means Clustering

Depthwise Separable Convolutional Neural Network for Confidential Information Analysis

Contact Info

Product

Resources

About