Home // SnT // News & E... // You May Also Like: SnT Researcher Enhances Recommender Privacy

You May Also Like: SnT Researcher Enhances Recommender Privacy

twitter linkedin facebook google+ email this page
Published on Thursday, 03 August 2017

Jun Wang of the Interdisciplinary Centre for Security, Reliability and Trust at the University of Luxembourg has been awarded the Best Student Paper Award at the 32nd International Conference on ICT Systems Security and Privacy Protection (IFIP SEC 2017).

The award was made in recognition of his work on privacy issues surrounding product recommendation systems, which are now such a major part of e-commerce websites. Such privacy issues are already well known, with significant attention paid to information leakage at the computing stage of producing these recommendations. However, Wang’s paper, ‘Differentially Private Neighborhood-based Recommender Systems’, addresses an overlooked problem: the recommendation results themselves can offer an attacker extensive sensitive information on a user.

“Recommender systems process vast amounts of historical browsing data, comparing it against a user’s data to predict their product preferences,” says Wang. “The issue we address is that malicious parties can then infer this data by analysing the recommendation results. From this inferred data, it doesn’t take much effort to gain personal information, such as political views, household income, gender etc. This makes privacy protection in this area crucial.”

Jun’s paper introduces a novel privacy-preserving solution to prevent information leakage from the recommendation results. This solution introduces ‘noise’ into the algorithm training process, limiting loss of privacy while still maintaining accurate product recommendations. Further, the solution quantifies the privacy loss inherent in using each recommender system, allowing users to make informed decisions as to whether they wish to use that system.

Jun Wang is a PhD student working under the direction of Prof. Peter Y. A. Ryan (SnT, APSIA Research Group). His co-supervisor is Qiang Tang (Luxembourg Institute of Science and Technology), and he is supported by a CORE (junior track) grant from the National Research Fund, Luxembourg.