Bounds for Pure Disjoint $(r,\delta)$-Quantum Locally Recoverable Codes

Technical Summary: Bounds for Pure Disjoint (r, δ)-Quantum Locally Recoverable Codes

Problem Statement
The paper addresses the fundamental limits of Quantum Locally Recoverable Codes (qLRCs), specifically focusing on the class of pure disjoint (r, δ)-qLRCs. While previous research has established bounds for qLRCs, particularly those constructed via the CSS framework or within the stabilizer formalism, there is a lack of general bounds for non-stabilizer codes that do not rely on connections to underlying classical codes. The authors aim to derive upper bounds on the code dimension $K$ for pure disjoint qLRCs using a weight-enumerator-based approach, without assuming a stabilizer structure. The specific setting involves codes where recovery sets are disjoint partitions of the physical qudits, allowing for the correction of up to $\delta-1$ erasures within any local recovery block of size $N = r + \delta - 1$.

Methodology
The authors develop a framework based on blockwise quantum weight enumerators and local Knill–Laflamme (KL) conditions.

1. Local Structure and KL Conditions:
   The paper first establishes that for disjoint qLRCs, the locality constraint induces a blockwise structure on the code space. The authors derive local KL conditions (Theorem 1), demonstrating that the correctability of erasures within a recovery block $J_\ell$ depends solely on the local code support $Q_\ell$. Specifically, an error is correctable on a block if and only if the local projectors satisfy specific scalar conditions on the error operators.

2. Blockwise Weight Enumerators:
   To capture the distribution of error weights across the disjoint recovery sets, the authors introduce two new types of enumerators:

   • Blockwise Shor–Laflamme (SL) enumerators ($A^{SL}_i, B^{SL}_i$): These track the weight distribution of errors across the $s$ recovery blocks.
   • Blockwise Unitary enumerators ($A^U_i, B^U_i$): These are defined via partial traces over subsets of qudits and relate to the SL enumerators through combinatorial identities (Corollary 1).

   The authors prove that these enumerators satisfy MacWilliams identities and inequalities analogous to the standard quantum case (Theorem 2 and Theorem 3). Crucially, they establish that for patterns of erasures that are correctable (defined by the set $T_{d,\delta}$), the blockwise enumerators satisfy a specific equality condition ($K B^U_i = A^U_i$).

3. Purity Assumption:
   The analysis focuses on pure codes, defined by the condition that the blockwise SL enumerators vanish for all non-zero correctable error patterns ($A^{SL}_i = 0$ for $i \in T_{d,\delta} \setminus \{0\}$). This assumption simplifies the relationship between the enumerators, allowing for the derivation of tighter bounds.

4. Derivation of Bounds:
   Using the properties of these enumerators and the Linear Programming (LP) method (adapted from classical coding theory and previous quantum works), the authors construct auxiliary polynomials based on multivariate Krawtchouk polynomials to derive upper bounds on the code dimension.

Key Contributions and Results

• Local KL Conditions for Disjoint qLRCs: The paper formulates necessary and sufficient conditions for local error correction in disjoint qLRCs, decoupling the local recovery capability from the global code structure.
• Blockwise Enumerators: The introduction of blockwise SL and unitary weight enumerators provides a new tool for analyzing qLRCs that respects the disjoint partition structure, bridging the gap between global quantum bounds and local recovery constraints.
• Singleton-like Bound (Theorem 4): For pure disjoint (r, δ)-qLRCs, the authors derive a Singleton-like bound:
  $$K \leq q^{n - 2(d-1) - 2\lfloor \frac{n-(d-1)}{N} \rfloor (\delta-1)}$$
  This bound is shown to be strictly stronger than the known bound for general disjoint (r, δ)-qLRCs (Equation 7) under the purity assumption. In the special case where $\delta=2$, it recovers the known stabilizer bound without requiring a stabilizer structure.
• Linear Programming Bound (Theorem 5 & 6): The authors derive an LP-based upper bound on the code dimension. By constructing specific polynomials with non-negative Krawtchouk coefficients, they obtain a bound (Equation 9) that is strictly tighter than relaxed versions of the Singleton-like bound and other known bounds in the literature.
• Non-Stabilizer Framework: A significant contribution is that these results are derived without assuming the code is a stabilizer code, offering a more general theoretical perspective on qLRCs.

Significance and Claims
The paper claims to provide a non-stabilizer, weight-enumerator-based approach to the study of pure disjoint (r, δ)-qLRCs. The significance lies in:

1. Strengthening Known Bounds: The derived Singleton-like bound improves upon existing bounds for disjoint qLRCs when the purity condition is met.
2. Generality: The methodology applies to general quantum codes, not just those with a stabilizer structure or those constructed via CSS codes.
3. Methodological Advancement: The use of blockwise enumerators allows for a precise accounting of how error weights are distributed across recovery sets, enabling the derivation of tighter constraints via the LP method.

The authors note that their LP bound coincides with the pure stabilizer bound in certain regimes and is at least as tight as the pure disjoint bound. They conclude that these results lay the groundwork for future extensions to impure codes and codes with overlapping recovery sets, mirroring developments in the classical setting.