… and strong privacy protection. To this end, we propose a DNN-based recommendation model
called … In [50], it is shown that a curious client in FL can infer not only membership, but also …

… to elaborate new recommendation systems taking privacy and … recommendation
system preserving both privacy and utility. The idea is to ensure protection against honest-but-curious …

… A curious or malicious user, or a coalition thereof, may attempt to make … privacy protection
into the computation itself, ensuring that the learned recommendations preserve privacy …

… More specifically, an attacker or curious recommendation server obtains user profiles and
can … However, these differential privacy based recommendation privacy protection schemes …

… We leverage uniform sampling to ensure differential privacy against curious users. Our
extensive evaluation demonstrates that XRec provides (1) recommendation quality similar to non-…

… -butcurious, which means server is curious in inferring … against attribute inference attack
would lead to suboptimal performance in both recommendation accuracy and privacy protection. …

… DP to check-in statistics before releasing them to an untrusted recommender system. That
method protects against both the recommender system, and its users, who may issue fictitious …

… -aware recommender system. Notably, privacy protection includes not only users’ privacy …
In APPLET, SP is curious-but-honest that is interested in uq’s recommendation results Rp at …

… proofs for the privacy protection of our proposed probabilistic differential privacy approach.
(… in geo-social networks: Proximity notification with untrusted service providers and curious …

… Besides preserving the privacy of the recommenders by … between trust establishment and
privacy protection. In this paper, we … This prevents a curious attacker from learning the values (r …