Secure and Efficient Computing on Private Data

Performing computations on private data, where the output of the computation is considered public, but the computations need to take place without compromising the privacy of the inputs (except for what the output itself reveals about the inputs) is an interesting problem nowadays. Application scenarios, where computing on private data is valuable, include ambient intelligence settings; advances in the field of medicine; home-healthcare applications; public emergency services’ support; autonomous driving applications; cloud computing services; and all applications that fall under the Internet of Things (IoT) class of applications. In this thesis we aim to treat the problems of computing on private data, and verifying performed computations on them, focusing specifically on medical applications, such as computations on DNA databases, which are extremely privacy sensitive. We focus on providing provably secure cryptographic solutions, allowing computations on private data, which at the same time are practically efficient. However, provable security comes at the cost of limited functionality, or decreased efficiency in terms of storage, computation, and communication of the proposed solutions. These factors can render promising privacy-preserving solutions inapplicable in practice, due to these increased costs, or insufficient functionality. We aim to develop fundamentally novel solutions that facilitate full functionality, while balancing all aforementioned trade-offs to achieve the highest efficiency, functionality, and user-friendliness, without compromising privacy. The prospective contributions of this thesis will be in the field of multiparty computation.