Biblio

Federated learning is a privacy-preserving learning system where participants locally update a shared model with their own training data. Despite the advantage that training data are not sent to a server, there is still a risk that a state-of-the-art model inversion attack, which may be conducted by the server, infers training data from the models updated by the participants, referred to as individual models. A solution to prevent such attacks is differential privacy, where each participant adds noise to the individual model before sending it to the server. Differential privacy, however, sacrifices the quality of the shared model in compensation for the fact that participants' training data are not leaked. This paper proposes a federated learning system that is resistant to model inversion attacks without sacrificing the quality of the shared model. The core idea is that each participant divides the individual model into model fragments, shuffles, and aggregates them to prevent adversaries from inferring training data. The other benefit of the proposed system is that the resulting shared model is identical to the shared model generated with the naive federated learning.

Location-based services, which provide services based on locations of consumers' interests, are becoming essential for our daily lives. Since the location of a consumer's interest contains private information, several studies propose location privacy protection mechanisms using an anonymizer, which sends queries specifying anonymous location sets, each of which contains k - 1 locations in addition to a location of a consumer's interest, to an LBS provider based on the k-anonymity principle. The anonymizer is, however, assumed to be trusted/honest, and hence it is a single point of failure in terms of privacy leakage. To address this privacy issue, this paper designs a semi-honest anonymizer to protect location privacy in NDN networks. This study first reveals that session anonymity and location anonymity must be achieved to protect location privacy with a semi-honest anonymizer. Session anonymity is to hide who specifies which anonymous location set and location anonymity is to hide a location of a consumer's interest in a crowd of locations. We next design an architecture to achieve session anonymity and an algorithm to generate anonymous location sets achieving location anonymity. Our evaluations show that the architecture incurs marginal overhead to achieve session anonymity and anonymous location sets generated by the algorithm sufficiently achieve location anonymity.