Visible to the public Cloud-Assisted IoT Systems Privacy--2019Q2Conflict Detection Enabled

Submitted by FengjunLi on Wed, 07/10/2019 - 2:18pm

PI(s), Co-PI(s), Researchers: Fengjun Li, Bo Luo

HARD PROBLEM(S) ADDRESSED

The goal of this project is to develop principles and methods to model privacy needs, threats, and protection mechanisms in cloud-assisted IoT systems. The work aims to address the hard problems of resilient architectures, security metrics as well as scalability and composability.

PUBLICATIONS

  • Anirudh Narasimman, Qiaozhi Wang, Fengjun Li, Dongwon Lee and Bo Luo, "Arcana: Enabling Private Posts on Public Microblog Platforms," in the 34rd International Information Security and Privacy Conference (IFIP SEC), Lisbon, Portugal, June 25-27, 2019.
  • Qiaozhi Wang, Hao Xue, Fengjun Li, Dongwon Lee, and Bo Luo, "#DontTweetThis: Scoring Private Information in Social Networks," in the 19th Privacy Enhancing Technologies Symposium (PETS), vol 4, Stockholm, Sweden, July 15-16, 2019.
  • Hao Xue, Qiaozhi Wang, Bo Luo, Hyunjin Seo, and Fengjun Li, "Content-Aware Trust Propagation Towards Online Review Spam Detection," to appear in ACM Journal of Data and Information Quality (JDIQ), 2019.

PUBLIC ACCOMPLISHMENT HIGHLIGHTS

  • We developed a privacy-preserving solution, called Arcana, to enable fine-grained access control for content sharing in online social networks. We presented this work in the 2019 IFIP International Information Security and Privacy Conference.

Using this tool, user can dynamically specify the dissemination boundary for the content they want to share and disseminate particular messages to designated group(s) of users (e.g., followers). We explored and extended a privacy enhancing technology, Ciphertext-Policy Attribute-based Encryption (CP-ABE), to implement social circle detection and private message encryption. The Arcana tool is lightweight and completely transparent to the underlying social networking platform, so it could be easily ported to other platforms to support fine-grained access control.

  • We developed a context-aware, text-based quantitative model for private information assessment. This work will be presented in the 2019 Privacy Enhancing Technologies Symposium (PETS).

A large amount of data, including those with private and sensitive information, have been posted online daily. However, there lacks any quantitative assessment on the privacy of the content, mainly due to its subjective nature. We are among the first to develop a computational scheme using deep neural networks to train prediction models and quantify privacy scores. With user's historical posts, topic preferences and social contexts, we generate personalized, context-aware privacy scores, which serve as the foundation for developing a user alerting mechanism to warn users when they attempt to disseminate sensitive information online.

  • We developed a social spam detection scheme, which leveraged the deviation of the aspect-specific individual opinions from the consensus opinions to quantify the trustworthiness of the users and the content they submit to online review platforms. Findings of this work will appear in ACM Journal of Data and Information Quality (JDIQ).

Online review systems become the target of individual and professional spammers, who insert deceptive content to manipulate the ratings and/or opinions. We model the influence of a user' opinion deviation on her trustworthiness into a deviation-based penalty and propagate this penalty to further measure the trustworthiness of the user's review content and review targets. The trust scores computed by our model are effective indicators of spammers and bots in online collaborative platforms.

COMMUNITY ENGAGEMENTS

  • We continued our bi-weekly Privacy Study Group in KU. Regular attendees include PIs and graduate students of our lablet, students from computer science and philosophy departments of KU, and researchers from Kansas State University.
  • Bo Luo served as a Grand Award Judge of 2019 Intel International Science and Engineering Fair, one of the world's largest STEM competition, May 12-17, 2019.
  • Bo Luo, presented "Arcana: Enabling Private Posts on Public Microblog Platforms" at IFIP International Information Security and Privacy Conference, Lisbon, Portugal, June 25, 2019.

EDUCATIONAL ADVANCES

We participated in regional and national collegian cyber-defense competitions.

Groups: