The Hardness of Nonlocal Quantum Computation

Loading...
Thumbnail Image

Journal Title

Journal ISSN

Volume Title

Publisher

University of Waterloo

Abstract

Position verification aims to certify a party’s location from the relativistic timing of its responses to a challenge. In the quantum setting its security reduces to a question about nonlocal quantum computation (NLQC). Every attack implements a bipartite unitary using pre-shared entanglement and a single simultaneous round of communication. Quantum position verification is information-theoretically insecure, since an attacker holding enough entanglement always succeeds, but it may remain complexity-theoretically secure if every attack requires an infeasible amount of entanglement. Establishing such security calls for lower bounds on the entanglement cost of NLQC, which are obstructed by connections to open problems in classical complexity. This thesis does not settle the security of position verification. Instead it studies the hardness of NLQC along two complementary routes. The first route develops a reduction calculus among NLQC tasks. We show that the classically controlled tasks f-route, f-measure, and CDQS are equivalent under constant-factor entanglement overhead, so that resource bounds, security properties, and upper bounds transfer freely among them. We then organize the coherently controlled tasks into a hierarchy, proving Cf-SWAP ⇒ Cf-PHASE ⇔Cf-PAULI; the weakest of these already recovers the classically controlled tasks. These reductions unify and propagate a number of bounds previously established in isolation. The second route studies unitary simulation directly. We bound how faithfully a fixed bipartite unitary can be reproduced by strategy classes weaker than full NLQC, from local operations, through entanglement-assisted local operations, to a single simultaneous round of quantum communication. The figure of merit is the fidelity of the simulated Choi state, which we control analytically through the spectral norm of a reduced Choi marginal. We apply the technique to three families of unitaries built from a Boolean function, the computational-basis (Rademacher), Weyl–Heisenberg-diagonal, and controlled-challenge families, the second containing maximally delocalizing perfect unitaries. In each family the technique ties hardness to an explicit property of the defining function (for the controlled family only down to a constant F ≥1 2 ), and identifies delocalization of the reduced Choi marginal as a structural feature that obstructs simulation by these restricted classes. The diagnostic is one-directional and certifies only a sufficient condition for hardness. It is not a hardness statement for full NLQC, in which even maximally delocalizing Clifford unitaries remain cheap to implement.

Description

Keywords

Citation

Endorsement

Review

Supplemented By

Referenced By