Description
We introduce Quantum Spectral Authentication (QSA), a primitive for verifying that a remote quantum endpoint still possesses a previously installed secret quantum resource such as a hidden state or state-preparation capability without revealing that secret [arXiv:2603.24868]. The core idea is to exploit spectral features of the planted
state under fresh public unitary challenges: each session, the verifier issues a new challenge unitary and the prover responds with a measurement outcome whose distribution is determined by the eigenstructure of the secret resource. Transcript-bound session material is derived from this exchange, providing explicit authentication without
exposing the underlying secret across repeated sessions.
We analyse the security of QSA against three classes of attack: eigenstate propagation across challenges, repeated-session leakage in which an adversary accumulates information over many transcripts, and direct online forgery. For each we characterise the conditions under which QSA remains secure and identify the parameters governing the information leakage rate.
For practical implementation on near-term hardware, we develop a symmetric verifier-driven unitary compiler compatible with low-depth quantum phase estimation. The key construction is a symmetric fast-power circuit in which both verifier and prover apply matched unitary sequences; simulations show this is substantially more noise tolerant than an asymmetric alternative, with the noise advantage growing with circuit depth. Small-instance experiments on the IBM \texttt{ibm_fez} device provide a hardware validation of the protocol at the scale accessible to current quantum processors.
QSA offers a plausible near-term authentication layer for control-plane applications in quantum computing. In addition, the framework connects naturally to quantum memory architectures in which a secret quantum state is maintained across authentication sessions, suggesting integration with emerging quantum network nodes.
| I am the presenting author | Yes |
|---|