arXiv:2609.23154
Quantum Hashing with QKD States
A. V. Vasiliev, I. G. Zinnatullin
Abstract
AI Summary
The paper asks whether quantum hashing can be implemented using the single-qubit states already available in QKD hardware, rather than state preparations requiring more advanced quantum engineering. It constructs a hash from a binary linear $[2k,n,d]$ error-correcting code: an $n$-bit input is encoded into $2k$ classical bits, with one half selecting computational or Hadamard bases and the other selecting basis vectors. The output is a product of $k$ BB84 states. The main general bound is $F(\psi_C)\leq 2^{-\lceil d/2\rceil}$, where $F$ is the maximum squared overlap between outputs for distinct inputs. The proof uses the code’s minimum distance: at least $\lceil d/2\rceil$ qubit positions differ, and each contributes at most $1/2$ to the product fidelity. An explicit code construction starts from a primitive binary BCH code, shortens it on information coordinates, and appends a zero coordinate if needed to obtain even length. For $n=32$ or $n\geq44$, it yields $d\geq2\lceil\log_2 n\rceil+1$, $k=\lceil n/2\rceil+O(\log^2 n)<n$, and $F(\psi_C)\leq1/(2n)$. The stated implementation motivation is that preparation and verification use operations available in existing QKD systems, requiring changes to their classical processing. The paper provides a construction and overlap analysis, not a hardware demonstration or noise analysis. It associates output compression with pre-image resistance but does not derive a separate quantitative pre-image-security bound.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.