Privacy-preserving distributed algorithm for sparse vector sum (Q983854)

Let \(B_0\) and \(B_1\) denote two local banks. Let each bank \(B_i\) has an \(n\)-dimensional input vector \(V_i\), where \(V_i[j]\) is the total amount of money bank \(B_i\) has loaned to customer \(j\). The purpose of this paper is to provide a privacy-preserving algorithm in which \(B_0\) and \(B_1\) do not reveal any information about their input vectors to each other except the output of algorithm. An analysis of the correctness, efficiency, and privacy guarantee of the proposed algorithm has also been provided. The computational indistinguishability follows from the semantic security of El-Gamal.