Privacy Preserving $n$-Party Scalar Product Protocol

Privacy-preserving machine learning enables the training of models on decentralized datasets without need to reveal data, both horizontal and vertically partitioned data. However, it relies specialized techniques algorithms perform necessary computations. The privacy preserving scalar product protocol, which dot vectors revealing them, is one popular example for its versatility. Unfortunately, solutions currently proposed in literature focus mainly two-party scenarios, even though scenarios with a higher number data parties are becoming more relevant. For when performing analyses that require counting samples fulfill certain criteria defined across various sites, such as calculating information gain at node decision tree. In this paper we propose generalization protocol an arbitrary parties, based existing method. Our solution recursive resolution smaller products. After describing our method, discuss potential scalability issues. Finally, describe guarantees identify any concerns, well comparing method original aspect.