arXiv:2609.39916
Distillation of N-Qubit Stabilizer States on a Star Network Topology
Theodore M. Mahaffey, Chaohan Cui, Saikat Guha, Murphy Yuezhen Niu
Abstract
We introduce an entanglement distillation protocol that utilizes an arbitrary $[[n,k,d]]$ stabilizer code to convert $n$ raw copies of an $N$-qubit Greenberger-Horne-Zeilinger (GHZ) state into $k$ logical copies in the presence of Pauli noise. Our explicit formulation of the scheme on a star network topology is efficiently scalable to arbitrary $N$, requires only local operations, and minimizes the impact of errors on idle physical qubits, enabling the distribution of high-fidelity entangled logical states between any number of quantum processors. We report the results of numerical experiments using the 5-qubit code, toric code, and $[[144,12,12]]$ bivariate bicyle code to protect against independent single-qubit depolarizing noise on all qubits. Our encoding scheme improves the final GHZ state fidelity relative to a single bare GHZ state sent over the same noisy channel for depolarizing rates up to or above $6 \%$ for all codes tested. Additionally, we use the stabilizer formalism to prove that our protocol can be applied to any $N$-qubit Calderbank-Shor-Steane (CSS) stabilizer state, from which the GHZ state emerges as a special case. This result explicitly establishes the working principle of previous stabilizer-based Bell pair and GHZ state distillation schemes, and generalizes such protocols beyond the traditionally considered states. We also show that some stabilizer codes can distill any stabilizer state with our protocol.
AI Summary
The paper studies code-based distribution of multipartite entanglement on a star network: when can local stabilizer measurements convert $n$ copies of an $N$-qubit state at a central hub into $k$ distributed logical copies protected by an $[[n,k,d]]$ code? The protocol measures code generators separately on each party’s block before transmission, sends the measurement outcomes classically, and performs local syndrome extraction and recovery after noisy distribution. Parties need not communicate with one another. Stabilizer-formalism proofs give sufficient compatibility conditions: codes admitting purely $X$-type logical $X$ and purely $Z$-type logical $Z$ representatives support any CSS stabilizer state, including GHZ states. Identical support for corresponding logical $X/Z$ representatives gives a stronger sufficient condition supporting all stabilizer states, with deterministic Pauli-frame adjustments where required. For independent Pauli channels, output GHZ fidelity is derived from single-party logical-error distributions, accounting for logical errors that collectively stabilize the target. Evaluation scales linearly with $N$ for heterogeneous channels, but remains exponential in $k$. For identical channels, the bound is $f_N\geq(1-p_{\mathrm{LE}})^N$, becoming tight at low logical-error rates. Exact five-qubit-code calculations and toric- and $[[144,12,12]]$ bivariate-bicycle-code decoding simulations benchmark depolarizing noise. The five-qubit code improves fidelity up to approximately $13.8\%$ depolarization for $N=3$–$100$; larger codes outperform a single transmitted bare GHZ state up to about $6\%$. Results assume perfect local operations and classical communication, with noise during transmission. Initial-state errors remain untested; heterogeneous noise models and hub measurement overhead are identified as further considerations.
Central claim may not hold: The claim that an arbitrary stabilizer code works is false: applying the protocol with the [[2,1,1]] code generated by Y1Y2 to two three-party GHZ states produces a product logical state even without noise. The demonstrated noise tolerance concerns distribution after noiseless encoding of ideal GHZ inputs, while robustness to errors already present in the raw input copies is explicitly left unestablished.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 22 pages, 4 figures