Privacy-preserving Recommender Systems Facilitated By The Machine Learning Approach

Doctoral thesis (2018)

Recommender systems, which play a critical role in e-business services, are closely linked to our daily life. For example, companies such as Youtube and Amazon are always trying to secure their profit by ... [more ▼]

Recommender systems, which play a critical role in e-business services, are closely linked to our daily life. For example, companies such as Youtube and Amazon are always trying to secure their profit by estimating personalized user preferences and recommending the most relevant items (e.g., products, news, etc.) to each user from a large number of candidates. State-of-the-art recommender systems are often built on top of collaborative filtering techniques, of which the accuracy performance relies on precisely modeling user-item interactions by analyzing massive user historical data, such as browsing history, purchasing records, locations and so on. Generally, more data can lead to more accurate estimations and more commercial strategies, as such, service providers have incentives to collect and use more user data. On the one hand, recommender systems bring more income to service providers and more convenience to users; on the other hand, the user data can be abused, arising immediate privacy risks to the public. Therefore, how to preserve privacy while enjoying recommendation services becomes an increasingly important topic to both the research community and commercial practitioners. The privacy concerns can be disparate when constructing recommender systems or providing recommendation services under different scenarios. One scenario is that, a service provider wishes to protect its data privacy from the inference attack, a technique aims to infer more information (e.g., whether a record is in or not) about a database, by analyzing statistical outputs; the other scenario is that, multiple users agree to jointly perform a recommendation task, but none of them is willing to share their private data with any other users. Security primitives, such as homomorphic encryption, secure multiparty computation, and differential privacy, are immediate candidates to address privacy concerns. A typical approach to build efficient and accurate privacy-preserving solutions is to improve the security primitives, and then apply them to existing recommendation algorithms. However, this approach often yields a solution far from the satisfactory-of-practice, as most users have a low tolerance to the latency-increase or accuracy-drop, regarding recommendation services. The PhD program explores machine learning aided approaches to build efficient privacy-preserving solutions for recommender systems. The results of each proposed solution demonstrate that machine learning can be a strong assistant for privacy-preserving, rather than only a troublemaker. [less ▲]

Differentially Private Neighborhood-based Recommender Systems

in IFIP Information Security & Privacy Conference (2017, May)

Privacy issues of recommender systems have become a hot topic for the society as such systems are appearing in every corner of our life. In contrast to the fact that many secure multi-party computation ... [more ▼]

Privacy issues of recommender systems have become a hot topic for the society as such systems are appearing in every corner of our life. In contrast to the fact that many secure multi-party computation protocols have been proposed to prevent information leakage in the process of recommendation computation, very little has been done to restrict the information leakage from the recommendation results. In this paper, we apply the differential privacy concept to neighborhood-based recommendation methods (NBMs) under a probabilistic framework. We first present a solution, by directly calibrating Laplace noise into the training process, to differential-privately find the maximum a posteriori parameters similarity. Then we connect differential privacy to NBMs by exploiting a recent observation that sampling from the scaled posterior distribution of a Bayesian model results in provably differentially private systems. Our experiments show that both solutions allow promising accuracy with a modest privacy budget, and the second solution yields better accuracy if the sampling asymptotically converges. We also compare our solutions to the recent differentially private matrix factorization (MF) recommender systems, and show that our solutions achieve better accuracy when the privacy budget is reasonably small. This is an interesting result because MF systems often offer better accuracy when differential privacy is not applied. [less ▲]

Privacy-preserving Friendship-based Recommender Systems

in IEEE Transactions on Dependable and Secure Computing (2016, November)

Privacy-preserving recommender systems have been an active research topic for many years. However, until today, it is still a challenge to design an efficient solution without involving a fully trusted ... [more ▼]

Privacy-preserving recommender systems have been an active research topic for many years. However, until today, it is still a challenge to design an efficient solution without involving a fully trusted third party or multiple semitrusted third parties. The key obstacle is the large underlying user populations (i.e. huge input size) in the systems. In this paper, we revisit the concept of friendship-based recommender systems, proposed by Jeckmans et al. and Tang and Wang. These solutions are very promising because recommendations are computed based on inputs from a very small subset of the overall user population (precisely, a user’s friends and some randomly chosen strangers). We first clarify the single prediction protocol and Top-n protocol by Tang and Wang, by correcting some flaws and improving the efficiency of the single prediction protocol. We then design a decentralized single protocol by getting rid of the semi-honest service provider. In order to validate the designed protocols, we crawl Twitter and construct two datasets (FMT and 10-FMT) which are equipped with auxiliary friendship information. Based on 10-FMT and MovieLens 100k dataset with simulated friendships, we show that even if our protocols use a very small subset of the datasets, their accuracy can still be equal to or better than some baseline algorithm. Based on these datasets, we further demonstrate that the outputs of our protocols leak very small amount of information of the inputs, and the leakage decreases when the input size increases. We finally show that he single prediction protocol is quite efficient but the Top-n is not. However, we observe that the efficiency of the Top-n protocol can be dramatically improved if we slightly relax the desired security guarantee. [less ▲]

Privacy-Preserving Context-Aware Recommender Systems: Analysis and New Solutions

in Computer Security - ESORICS 2015 - 20th European Symposium on Research in Computer Security (2015, September)

Nowadays, recommender systems have become an indispens- able part of our daily life and provide personalized services for almost everything. However, nothing is for free – such systems have also upset the ... [more ▼]

Nowadays, recommender systems have become an indispens- able part of our daily life and provide personalized services for almost everything. However, nothing is for free – such systems have also upset the society with severe privacy concerns because they accumulate a lot of personal information in order to provide recommendations. In this work, we construct privacy-preserving recommendation protocols by incorpo- rating cryptographic techniques and the inherent data characteristics in recommender systems. We first revisit the protocols by Jeckmans et al. and show a number of security issues. Then, we propose two privacy- preserving protocols, which compute predicted ratings for a user based on inputs from both the user’s friends and a set of randomly chosen strangers. A user has the flexibility to retrieve either a predicted rating for an unrated item or the Top-N unrated items. The proposed protocols prevent information leakage from both protocol executions and the pro- tocol outputs. Finally, we use the well-known MovieLens 100k dataset to evaluate the performances for different parameter sizes. [less ▲]