Asymmetric Entanglement-Assisted Quantum Locally Recoverable Codes

AI summary

Overview

Research area: Quantum error correction — specifically quantum locally recoverable codes (qLRCs), entanglement-assisted quantum codes (EAQECCs), asymmetric quantum codes, and the classical coding theory (LRCs, Tamo–Barg codes, cyclic codes) used to construct them.

Technical level: Advanced. The paper is written in the language of finite fields, dual codes, stabilizer formalism, and coding-theoretic bounds; the introduction and the plain statements of the main results are accessible to an intermediate reader, but the theorems themselves require graduate-level algebraic coding theory.

Scope (one sentence): The paper introduces and analyzes asymmetric entanglement-assisted quantum locally recoverable codes (AEA-qLRCs), in which bit-flip (X-type) and phase-flip (Z-type) errors are recoverable from recovery sets of size at most r_X+1 and r_Z+1 respectively, and derives stabilizer criteria, a CSS-type construction, five families of bounds, exact Singleton-optimality conditions, and explicit Tamo–Barg and cyclic constructions.

What This Paper Is About

Quantum locally recoverable codes let an erased qudit be rebuilt by accessing only a small number of surviving qudits, which matters for distributed quantum storage and quantum memories. Every existing qLRC and entanglement-assisted qLRC construction assumes a single locality parameter r that governs recovery from both X-type and Z-type errors, even though physical noise is rarely symmetric between the two error types. This paper asks whether entanglement assistance, asymmetric error correction, and local recoverability can be achieved simultaneously, and answers yes by defining AEA-qLRCs with independent localities {r_X, r_Z}, characterizing them in the stabilizer formalism, and supplying bounds and explicit constructions.

Key Contributions

1. Definition and stabilizer characterization of AEA-qLRCs. The authors define a code with locality {r_X, r_Z} in which X-type and Z-type errors at each coordinate are independently recoverable from sets of size at most r_X+1 and r_Z+1 using shared entanglement. Theorem 7 proves that a pair of stabilizer generators which reduce, on the encoder side, to a pure X-type and a pure Z-type operator at coordinate i, with supports of size at most r_X+1 and r_Z+1 respectively, suffices for this property.

2. A CSS-type construction (Proposition 8). Any two classical codes C_1, C_2 whose duals admit, at every coordinate, parity checks of weight at most r_Z+1 and r_X+1 respectively yield an AEA-qLRC with locality {r_X, r_Z}. The two localities are therefore controlled independently by the two classical codes.

3. Five families of bounds. Theorem 9 gives a general trade-off bounding d_X + d_Z, the logical dimension κ, and the length n via the optimal-parameter functions n_opt^(q), k_opt^(q), d_opt^(q) of classical LRCs. Applying known classical LRC bounds produces Singleton-, Cadambe–Mazumdar-, Plotkin-, sphere-packing-, and Griesmer-like bounds for CSS-type AEA-qLRCs (Corollaries 10 and 11).

4. Exact optimality characterizations and explicit constructions. Theorems 12, 14, and 15 characterize when the Singleton-like bound is met with equality, restate the criterion in terms of s_1 = dim(C_1 ∩ C_2^⊥) and s_2 = dim(C_2 ∩ C_1^⊥), and identify optimal maximally entangled codes with Linear Complementary Pairs of optimal LRCs. Proposition 16 and Theorems 17–18 instantiate the framework on Tamo–Barg codes (disjoint- and overlapping-support regimes), and Lemma 21 with Theorems 22–23 do the same for cyclic LRCs.

Main Findings

• Asymmetric locality is a strict generalization of the symmetric theory. Setting r_X = r_Z = r throughout recovers exactly the stabilizer criterion, CSS construction, and bounds of the concurrent symmetric EA-qLRC work (Li et al. 2026b; Kumar and Bandi 2026), so the results are presented as a strict asymmetric generalization rather than an independent development.

• Recovery-set sizes are separate, not shared. X-type errors are recovered from sets of size at most r_X+1 and Z-type errors from sets of size at most r_Z+1, whereas the prior symmetric framework used one parameter r for both.

• A single trade-off theorem generates all five bounds. Theorem 9 bounds d_X + d_Z, κ, and n via the classical LRC optimal-parameter functions n_opt^(q), k_opt^(q), d_opt^(q); Corollaries 10 and 11 specialize it to the Singleton-, Cadambe–Mazumdar-, Plotkin-, sphere-packing-, and Griesmer-like bounds.

• Singleton optimality has a sharp characterization. For pure CSS-type AEA-qLRCs built from a pair of classical LRCs, equality in the Singleton-like bound holds if and only if C_1 and C_2 are optimal LRCs with d_Z = d_X and matching locality-ceiling conditions (Theorem 12). The condition is restated in terms of s_1 and s_2 in Theorem 14.

• Maximally entangled optima require equal distances. Optimal maximally entangled AEA-qLRCs correspond exactly to Linear Complementary Pairs of optimal LRCs, and this requires d_X = d_Z (Theorem 15) — a constraint that the symmetric counterpart of the result did not carry.

• Tamo–Barg constructions allow independently tunable localities. Proposition 16 explicitly evaluates the relevant dual-intersection sizes for Tamo–Barg codes, yielding pure AEA-qLRC constructions in both the disjoint-support and overlapping-support regimes (Theorems 17 and 18), with r_X and r_Z tunable independently.

• Cyclic LRCs give a second construction family. A support-alignment lemma (Lemma 21) yields a general cyclic construction (Theorem 22), and Linear Complementary Pairs of optimal cyclic LRCs yield an optimal maximally entangled family (Theorem 23).

• A known Singleton-optimality obstruction is inherited from the symmetric setting. The authors' prior symmetric work established that Tamo–Barg-based constructions attain the Singleton-like bound only in the vacuous regime k ≤ r; the asymmetric constructions here are presented as tunable rather than as Singleton-attaining in general.

• Not reported in the available content. The truncated text does not contain explicit numerical code tables, worked parameter examples, asymptotic tightness comparisons for the new asymmetric bounds, or Gilbert–Varshamov-type achievability bounds for the asymmetric setting.

Methodology in Plain English

The authors work entirely within the stabilizer formalism of quantum error correction. In a stabilizer code, the code space is fixed by a set of commuting Pauli operators; to check whether a coordinate can be locally repaired, one looks for stabilizer generators that act nontrivially on that coordinate in a controlled way. For asymmetric locality, the authors require two such generators: one that acts as a pure bit-flip-type operator at coordinate i (with small support), and one that acts as a pure phase-flip-type operator at coordinate i (also with small support). The first handles X-type errors from a set of size at most r_X+1, the second handles Z-type errors from a set of size at most r_Z+1. Adding pre-shared entangled pairs on the decoder side relaxes the requirement that the encoder-side operators commute, which is what makes entanglement assistance compatible with the local-recovery requirement.

For explicit construction, the authors reduce the quantum problem to a classical one: build the code from two classical linear codes C_1 and C_2, and require that their duals each contain a low-weight parity check covering every coordinate — weight at most r_Z+1 for C_1 and at most r_X+1 for C_2. This CSS-style reduction means all the machinery of classical locally recoverable codes (their bounds, Tamo–Barg codes, cyclic codes) can be imported directly. The number of entangled pairs consumed is c = rank(H_1 H_2^T), and the X- and Z-distances are computed from set differences involving C_1, C_2 and their duals. Bounds are then obtained by plugging the classical LRC Singleton, Cadambe–Mazumdar, Plotkin, sphere-packing, and Griesmer bounds into the general trade-off theorem, and optimality is analyzed by asking exactly when those inequalities become equalities.

Why This Matters

Impact on research. The paper closes a specific gap at the intersection of three active lines of work: entanglement-assisted qLRCs (single locality parameter), asymmetric qLRCs (two localities, no entanglement), and asymmetric EAQECCs (asymmetric errors, no locality). It provides the first unified framework in which all three are present at once, supplies the converse bounds needed to judge whether future constructions are good, and pins down exactly which code pairs can be Singleton-optimal — information that constrains future construction attempts rather than merely describing existing ones.

Real-world applications.

• Quantum distributed storage: stored qudits spread across nodes, where a failed node is repaired by contacting only a few neighbors, and where the repair cost for phase-flip protection can be budgeted differently from bit-flip protection.
• Quantum memories with biased noise: hardware in which Z-type (phase) errors dominate X-type errors, cited in the paper via Steane (1996), can be protected with r_Z larger than r_X instead of paying a symmetric cost.
• Quantum repeaters and networked quantum computing: pre-shared entanglement is already a resource in these settings, so entanglement-assisted recovery does not require a new primitive.
• Fault-tolerant quantum architectures: reducing the number of qudits touched during a recovery operation reduces the circuit depth and the opportunity for new errors during repair.

Industry relevance. Providers of quantum cloud services and builders of quantum memory and networking hardware are the natural consumers: the asymmetric model matches the actual error bias of many physical platforms better than a symmetric locality parameter, and the entanglement-assisted model matches the fact that entanglement distribution is a designed-in feature of networked quantum systems rather than an exotic add-on.

Future Directions

• Gilbert–Varshamov-type achievability for the asymmetric case. The authors' prior symmetric framework included two Gilbert–Varshamov-like achievability bounds via parity-check augmentation and concatenation; the contribution list here does not include an asymmetric analogue, leaving the question of how close the new converse bounds are to achievable open.

• Tightness and asymptotic analysis of the five new bounds. The symmetric companion work analyzed the relative asymptotic and regime-specific tightness of its five bounds; the corresponding analysis for the asymmetric bounds, and the question of which of r_X or r_Z dominates in each regime, is not addressed in the available content.

• Beyond CSS, and beyond the pure-code case. The construction and optimality results are stated for CSS-type AEA-qLRCs built from Proposition 8; non-CSS stabilizer constructions, Hermitian constructions over F_{q^2}, and impure codes that beat the pure-code Singleton condition all remain to be treated.

• Explicit parameter tables and comparisons. The paper states the Tamo–Barg and cyclic constructions as theorems but the available content reports no numerical code tables; producing explicit [[n, κ, {d_X, d_Z}; c]]_q families and comparing c (the entanglement cost) against symmetric baselines is the natural next empirical step, as is extending the constructions from the fields used in the Tamo–Barg definition to arbitrary finite fields.

Target Audience

Researchers in quantum error correction and quantum information theory who work on locality-constrained codes, stabilizer constructions, or entanglement-assisted codes; coding theorists interested in importing classical LRC bounds (Singleton, Griesmer, Plotkin, sphere-packing, Cadambe–Mazumdar) into the quantum setting; and graduate students or advanced practitioners who already know the CSS construction and classical locally recoverable codes and want a self-contained treatment of the asymmetric, entanglement-assisted generalization. Readers without a background in finite-field coding theory will find the preliminaries section necessary but the main theorems demanding.

Authors’ abstract

Locally recoverable codes (LRCs) permit reconstruction of a lost coordinate from a small subset of the remaining coordinates, a property that is a fundamental coding technique for distributed storage and, more recently, for fault-tolerant quantum memories and distributed storage systems. Existing constructions of quantum locally recoverable codes (qLRCs) and their entanglement-assisted generalizations (EA-qLRCs) impose a single locality parameter $r$ governing recovery from both bit-flip ($X$-type) and phase-flip ($Z$-type) errors, even though physical noise is rarely symmetric between the two. We introduce entanglement-assisted quantum locally recoverable codes (EA-qLRC) with \emph{asymmetric locality} $\{r_X,r_Z\}$, in which $X$-type and $Z$-type errors are corrected from recovery sets of size at most $r_X+1$ and $r_Z+1$, respectively. We extend a stabilizer sufficient condition for this property and, in particular, the CSS framework, where asymmetric locality corresponds to separate locality conditions on the two classical codes $\mathcal C_1$ and $\mathcal C_2$ underlying the construction. From this characterization we derive Singleton-, Griesmer-, Plotkin-, sphere-packing-, and Cadambe\textendash Mazumdar-like bounds for CSS-type asymmetric EA-qLRCs (AEA-qLRCs) built from Proposition~\ref{CSS-EA-qLRC-Asym}, and we characterize exactly when a pure CSS-type AEA-qLRC built from a pair of classical LRCs attains the Singleton-like bound with equality.