Achieves 0.94 Fidelity: LiDMaS Advances Fault-Tolerant GKP Photonic Qubit Injection

Why It Matters

The results demonstrate that high‑fidelity magic‑state preparation is feasible with modest hardware, informing near‑term photonic quantum computer designs and reducing the squeezing burden for fault‑tolerant operation.

Key Takeaways

• Success probability exceeds 94% across all tested parameters
• Average injection overhead stays near unity
• Logical fidelity reaches 0.77‑0.80 after surface‑code protection
• Squeezing level dominates fidelity, loss impact minimal
• Framework maps squeezing to logical dephasing, loss to heralded erasure

Pulse Analysis

Fault‑tolerant quantum computing hinges on reliable magic‑state injection, yet photonic platforms grapple with finite GKP squeezing and photon loss. LiDMaS sidesteps costly wavefunction simulations by representing encoded qubits with 2 × 2 density matrices, enabling rapid exploration of architectural trade‑offs. This abstraction captures essential error channels—dephasing from squeezing, depolarizing noise, and heralded erasure from loss—while remaining computationally lightweight, a crucial advantage for designers evaluating large parameter spaces.

The study’s repeat‑until‑success (RUS) injection protocol, combined with an outer surface‑code layer, delivered success probabilities consistently above 94% and kept the average injection overhead close to one, a rare achievement in photonic schemes. Logical fidelities after surface‑code protection settled between 0.77 and 0.80, showing weak sensitivity to moderate loss (1‑3%) but a strong, monotonic dependence on squeezing levels from 8 dB to 16 dB. Phase‑boundary analyses pinpointed the minimum squeezing required to meet a 95% success threshold and a 0.79 fidelity target, offering concrete benchmarks for experimental teams.

These findings reshape the roadmap for scalable photonic quantum computers. By quantifying the squeezing budget needed for fault‑tolerant operation, LiDMaS equips architects with actionable design margins, potentially accelerating hardware development cycles. The framework’s modular noise mapping also invites extensions to more complex error models, paving the way for holistic, architecture‑level optimization across diverse quantum platforms. Confidence in photonic fault tolerance thus moves from theoretical promise toward engineering reality.