arXiv:2609.26898
The logical set of Gaussian states under the stabilizer subsystem decomposition
Santiago Zamora, Kaustav Chatterjee, Ulrik Lund Andersen, Thiago Lucena de Macedo Guedes, Jonatan Bohr Brask, A. de Oliveira Junior, Rafael Chaves
Abstract
Universal quantum computation requires non-Gaussianity in continuous-variable systems and non-stabilizerness in discrete-variable systems. Yet, mapping a Gaussian state into the Gottesman-Kitaev-Preskill logical subspace can yield a logical qubit with non-stabilizer resources. To understand how the continuous-variable structure determines logical resources, we characterize the image of single-mode Gaussian states under the corresponding extraction channel, known as the stabilizer subsystem decomposition. Specifically, we derive series expressions for the logical Bloch vector and characterize the resulting reachable set, enabling an analytical treatment of the robustness of magic of the logical states. We then turn this geometric framework into a certification tool: by mapping a discrete-variable qubit witness to a continuous-variable squeezing witness, we obtain lower bounds on the squeezing of prepared states. We also show that ensembles of Gaussian states can violate both stabilizer and classical bounds. Violation of the stabilizer bound certifies the relational resource known as "set-magic," while violation of the classical bound enables the certification of cryptographic randomness at moderate squeezing.
AI Summary
Which logical qubit states can single-mode Gaussian states produce after ideal GKP error correction and syndrome-discarding logical extraction? The paper studies this stabilizer subsystem decomposition (SSD), relating Gaussian covariance and displacement parameters to logical non-stabilizerness. It derives exact, absolutely convergent series for all three logical Bloch components and analytically characterizes the $XZ$ projection. This projection is convex, with a strictly concave boundary generated by pure, axis-aligned squeezed states. Writing $A(t)=\frac4\pi\sum_{m\ge0}\frac{(-1)^m}{2m+1}e^{-\pi(2m+1)^2t/2}$, the positive-quadrant boundary is $(A(t),A(1/(4t)))$. Theta-function identities establish displacement extremality and monotonicity; a modular identity establishes boundary concavity. The vacuum maximizes $|X|+|Z|$, giving $2A(1/2)\approx1.160314$. A leading-harmonic estimate gives robustness $\approx1.302$ at squeezing $r\approx0.262$, consistent with a previously reported numerical maximum of $1.303$. Explicit Gaussian families approach all six stabilizer vertices; their mixtures approximate every stabilizer-polytope point with one-sided Euclidean error at most $(2\sqrt2/\pi)e^{-r}(1+o(1))$. The component bound $\max(|X|,|Z|)\le A(e^{-2r_0}/2)$ provides a tight Gaussian squeezing witness. Prepare-and-measure constructions violate stabilizer bounds, certifying set-magic, while numerical optimization and semidefinite relaxations certify approximately $0.18$ randomness bits at $r_0\gtrsim1.5$. The full three-dimensional boundary and global magic optimum remain unresolved analytically. Certification assumes Gaussian inputs and faithful ideal SSD; SSD itself supplies a non-Gaussian resource. Randomness minimization samples near-degenerate local optima rather than exhaustively treating all optimizers. Finite-energy GKP extraction, realistic noise, and multimode extensions remain open.
Error found: The numerical ensemble witness benchmarks are not certified global upper bounds, so exceeding them alone does not rigorously certify squeezing above the stated limit. The reported worst-case randomness is minimized over a sampled collection of numerical optima rather than the entire maximizing set, leaving its universal worst-case interpretation unproved.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 24 pages, 2 figures