Grant: Discovery Grants Program - Individual

When it comes to personal information, a tension exists between society's need to provide efficient and effective services, and an individual's right to privacy. Many services benefit from access to detailed data about individuals when allocating their limited resources. For example, greater access to patient information allows medical researchers to develop new and more personalized treatments. At the same time, a range of policies and laws have emerged to protect individual privacy by restricting the sharing of sensitive personal information. Any methodology that could simultaneously meet the analytical needs of service providers while protecting individual privacy, therefore, would be useful to a society.*Secure multi-party computation confronts the problem of how two or more parties can communicate to answer questions of mutual importance, without the participants revealing more to each other than what is revealed by the result itself. Using cryptographic techniques, this approach allows individuals and organizations to combine the analytical power of their personal and private information without ever sharing it directly. This opens the door to fundamentally new possibilities for answering important questions that rely on access to personal data.*We propose a research program exploring new applications of secure computation to questions of practical, real-world importance. The primary area of focus will be to develop secure computation techniques in the field of healthcare. Other key areas of focus will include secure online voting and digital currency. Our previous research experience in these focus areas leads us to the following methodology: (1) consult with stakeholders of services in these application areas to identify questions of analytical importance involving personal and private data; (2) define a set of security requirements and develop a secure multi-party computation protocol to realize them; (3) analyze the security properties of the protocol to ensure it meets the requirements, and (4) prototype the protocol in software to evaluate its computational efficiency, and hence its feasibility.