Quantum Garbled Circuits

We present a garbling scheme for quantum circuits, thus achieving a decomposable randomized encoding scheme for quantum computation. Specifically, we show how to compute an encoding of a given quantum circuit and quantum input, from which it is possible to derive the output of the computation and nothing else. In the classical setting, garbled circuits (and randomized encodings in general) are a versatile cryptographic tool with many applications such as secure multiparty computation, delegated computation, depth-reduction of cryptographic primitives, complexity lower-bounds, and more. However, a quantum analogue for garbling general circuits was not known prior to this work. We hope that our quantum randomized encoding scheme can similarly be useful for applications in quantum computing and cryptography. To illustrate the usefulness of quantum randomized encoding, we use it to design a conceptually-simple zero-knowledge (ZK) proof system for the complexity class $\mathbf{QMA}$. Our protocol has the so-called $\Sigma$ format with a single-bit challenge, and allows the inputs to be delayed to the last round. The only previously-known ZK $\Sigma$-protocol for $\mathbf{QMA}$ is due to Broadbent and Grilo (FOCS 2020), which does not have the aforementioned properties.