Abstract
Consider a scenario where a computationally weak party, the client, wishes to dele-gate his/her desired quantum computation to a more powerful party, the server. While delegating expensive tasks to quantum computers in the cloud seems promising, it raises an inevitable concern: How can a client ensure the privacy of their computa-tion as well as the input data and at the same time verify that the output is indeed correct? In this dissertation, I investigate the task of the secure delegation of quantum computation or blind quantum computation. To do this, I introduce secure protocols in the client-server setting and analyze them under the information-theoretic security approach. In the first part of this dissertation, I propose a blind quantum computing scheme for a client with no quantum capability. The protocol allows a classical client to com-pute only the measurement angles to delegate the computation to a remote server via multiple rounds of interaction. I analyze the security of the protocol in both the stand-alone and composable security framework. Importantly, the security relies on the flow of information, a notable feature of measurement-based quantum computing. Further-more, I observe that measurements in the XY-plane on cluster states are sufficient to implement universal quantum computation. In the second part of this dissertation, I make advances on the existing blind and verifiable quantum computing schemes for clients with limited quantum computa-tional abilities. I introduce a protocol which requires a single round of interaction between the client and the server by exploiting a recent construction for secure (prob-abilistic) one-time programs. Finally, I generalize the delegated quantum computing scenario to multiple clients. In particular, I show that the proposed multi-client proto-col is both blind and verifiable, when either the client(s) or server is malicious.