Privacy-Preserving Cooperative Computations

Jan 31, 2001

Abstract

The Internet has triggered tremendous opportunities for cooperative computation, where people are cooperating with each other to conduct computation tasks based on the inputs they each supplies. These computations could occur between completely trusted partners, between partially trusted partners, or between mutually untrusted competitors. For example, customers might send to a remote database the queries that contain private information, two competing financial organizations might jointly invest in projects that must satisfy both organizations' private and valuable constraints, and so on. Usually, to conduct these computations, one must know inputs from all the participants; however if nobody can be trusted enough to know all the inputs, privacy will become a primary concern. Therefore, the problem is "how can multiple parties cooperatively conduct a computation without disclosing their private inputs to the other parties".

This problem is referred to as Secure Multi-party Computation Problem (SMC) in the literature, but the research in the SMC area has been focusing on only a limited set of problems. We have identified a number of new SMC problems for a spectrum of computation domains. In this talk, I will describe these new problems, including privacy-preserving database query, privacy-preserving data mining, privacy-preserving intrusion detection, privacy-preserving scientific computations, privacy-preserving statistical analysis, and privacy-preserving geometric computations. We will also discuss their potential applications. We have developed protocols to solve some of the above problems, although not all of them. I will present one to two of our protocols in this talk to demonstrate how this type of problem could be solved. Joint work with Professor Mikhail Atallah.