We investigate the possibility of having someone carry out the work of executing a function for you, but without letting him learn anything about your input. Say Alice wants Bob to compute some known function f upon her input x, but wants to prevent Bob from learning anything about x. The situation arises for instance if client Alice has limited computational resources in comparison with mistrusted server Bob, or if x is an inherently mobile piece of data. Could there be a protocol whereby Bob is forced to compute ,f(x)blindly, i.e. without observing x? We provide such a blind computation protocol for the class of functions which admit an efficient procedure to generate random input–output pairs, e.g. factorization. The cheat-sensitive security achieved relies only upon quantum theory being true. The security analysis carried out assumes the eavesdropper performs individual attacks.
We investigate the possibility of having someone carry out the work of executing a function for you, but without letting him learn anything about your input. Say Alice wants Bob to compute some known function f upon her input x, but wants to prevent Bob from learning anything about x. The situation arises for instance if client Alice has limited computational resources in comparison with mistrusted server Bob, or if x is an inherently mobile piece of data. Could there be a protocol whereby Bob is forced to compute ,f(x)blindly, i.e. without observing x? We provide such a blind computation protocol for the class of functions which admit an efficient procedure to generate random input–output pairs, e.g. factorization. The cheat-sensitive security achieved relies only upon quantum theory being true. The security analysis carried out assumes the eavesdropper performs individual attacks.
Suppose Alice wants to perform some computation that could be done quickly on a quantum computer, but she cannot do universal quantum computation. Bob can do universal quantum computation and claims he is willing to help, but Alice wants to be sure that Bob cannot learn her input, the result of her calculation, or even the function she is trying to compute. We describe an efficient protocol by which Bob can help Alice perform the computation, but there is no way for him to learn anything about it. Furthermore, we show how Alice can efficiently detect whether Bob is honestly helping her or if he is introducing errors.
Blind quantum computation is a new secure quantum computing protocol which enables Alice (who does not have sufficient quantum technology) to delegate her quantum computation to Bob (who has a full-fledged quantum computer) in such a way that Bob cannot learn anything about Alice's input, output, and algorithm. In previous protocols, Alice needs to have a device which generates quantum states, such as single-photon states. Here we propose another type of blind computing protocol where Alice does only measurements, such as the polarization measurements with a threshold detector. In several experimental setups, such as optical systems, the measurement of a state is much easier than the generation of a single-qubit state. Therefore our protocols ease Alice's burden. Furthermore, the security of our protocol is based on the no-signaling principle, which is more fundamental than quantum physics. Finally, our protocols are device independent in the sense that Alice does not need to trust her measurement device in order to guarantee the security.
Suppose Alice wants to perform some computation that could be done quickly on a quantum computer, but she cannot do universal quantum computation. Bob can do universal quantum computation and claims he is willing to help, but Alice wants to be sure that Bob cannot learn her input, the result of her calculation, or perhaps even the function she is trying to compute. We describe a simple, efficient protocol by which Bob can help Alice perform the computation, but there is no way for him to learn anything about it. We also discuss techniques for Alice to detect whether Bob is honestly helping her or if he is introducing errors.
Abstract Blind quantum computing enables quantum-limited clients to securely access untrusted quantum servers. This paper proposes a three-party verifiable blind quantum computing protocol with separated input and function ownership, where Alice holds the private input, Bob holds the target function, and Charlie performs the delegated quantum computation, with the result released only to Alice. Alice and Bob first establish a shared key using BB84. Under the measurement-based quantum computing framework, a secure two-party computation protocol is then used to jointly generate adaptive blinded measurement instructions, eliminating the need for Bob to transmit quantum states to Charlie. A hidden dummy-trap mechanism is further incorporated to verify malicious server behavior. Under the honest-but-curious client model without client-server collusion, the protocol protects input and function privacy while ensuring computation blindness, shared-key security, and verifiability. Most quantum operations are performed by Charlie, while the clients require only limited quantum capabilities.
Blind quantum computation is a new secure quantum computing protocol where a client, who does not have enough quantum technologies at her disposal, can delegate her quantum computation to a server, who has a fully fledged quantum computer, in such a way that the server cannot learn anything about the client's input, output, and program. If the client interacts with only a single server, the client has to have some minimum quantum power, such as the ability of emitting randomly rotated single-qubit states or the ability of measuring states. If the client interacts with two servers who share Bell pairs but cannot communicate with each other, the client can be completely classical. For such a double-server scheme, two servers have to share clean Bell pairs, and therefore the entanglement distillation is necessary in a realistic noisy environment. In this Letter, we show that it is possible to perform entanglement distillation in the double-server scheme without degrading the security of blind quantum computing.
A user, Alice, wants to get server Bob to implement a quantum computation for her. However, she wants to leave him blind to what she's doing. What are the minimal communication resources Alice must use in order to achieve information-theoretic security? In this paper, we consider a single step of the protocol, where Alice conveys to Bob whether or not he should implement a specific gate. We use an entropy-bounding technique to quantify the minimum number of qubits that Alice must send so that Bob cannot learn anything about the gate being implemented. We provide a protocol that saturates this bound. In this optimal protocol, the states that Alice sends may be entangled. For Clifford gates, we prove that it is sufficient for Alice to send separable states.