The Quantum Hamming Bound in Arbitrary Local Dimension

Abstract: The quantum Hamming bound is the finite-length sphere-packing count for exact quantum error correction: the code dimension times the number of correctable local error patterns must fit inside the ambient Hilbert space. For nondegenerate codes this follows from disjoint error spheres. Degeneracy is the only obstruction, because distinct physical errors can coincide on the code subspace and turn sphere packing into an overcount. The central finite-length question has been whether this overcount can ever invalidate the Hamming inequality. Earlier linear-programming, asymptotic, and structural results left a pointwise finite-length problem for arbitrary exact subspace codes. Writing $Q=q^2$ for the Hamming-scheme alphabet, the nonbinary range begins at $Q=9$; here we prove the bound for every $q\ge3$, while the binary endpoint is governed by a distinct $Q=4$ charging geometry. For every nontrivial exact subspace code in this range, any possible violation reduces to a two-center Hamming-ball intersection inequality normalized by the Lloyd response. For $q\ge4$, the Lloyd-square linear program has a uniform half-gap after reduction to alphabet $Q\ge16$ and critical length $n=4t+1$. Qutrits form the boundary: the half-gap disappears, but the bridge is handled by a quadratic-filtered Lloyd square, an exact coefficient-certificate reduction, and a Stein-tangent positivity argument. Thus degeneracy may merge error sectors, but not enough to beat the Hamming count. This proves the nonbinary part of the finite-length quantum Hamming bound; together with the independent binary endpoint theorem, it gives the result in arbitrary local dimension.

• The paper establishes that every exact quantum subspace code (n, K, d)_q with K>1 satisfies the quantum Hamming bound for all q≥2.
• It employs a linear programming method with a two-center Hamming-ball approach and quadratic filters to control overlaps in degenerate codes.
• Numerical and algebraic proofs confirm finite-length guarantees crucial for designing robust quantum memory and communication systems.

Summary of "The Quantum Hamming Bound in Arbitrary Local Dimension" (2606.22538)

Context and Finite-Length Quantum Coding Theory

This paper resolves a longstanding finite-length question in quantum error correction: whether the quantum Hamming bound holds for exact subspace codes in arbitrary local dimension, including all degenerate and nonadditive codes. The quantum Hamming bound is a sphere-packing inequality: the code dimension $K$ times the number of correctable local error patterns $V_t(n)$ must fit inside the ambient Hilbert space $q^n$. While for nondegenerate codes this follows from disjointness of error spheres, degeneracy introduces potential sector overlaps where distinct physical errors coincide on the code subspace. The central problem has thus been whether these overlaps can ever be large enough to invalidate the Hamming bound.

Previous results established the bound asymptotically and for specific code families (stabilizer, additive) or for sufficiently large lengths and local dimensions, but no general finite-length theorem covered the full nonbinary range (qutrits and higher). The present work closes this gap and proves: every exact quantum subspace code $(n,K,d)_q$ with $K>1$ satisfies the quantum Hamming bound for all $q\geq2$.

Methodology and Technical Framework

The proof employs a linear programming (LP) approach, inspired by the Li-Xing quantum LP bound. The analysis centers on the Q-ary Hamming scheme, where $Q=q^2$ is the alphabet size of the local error labels. Rather than relying on single-center error balls, the proof uses a two-center Hamming-ball intersection, normalized by the Lloyd response. The challenge is to control the ratio of overlapping sectors, especially in degenerate codes where the usual disjoint sphere picture breaks down.

For $q\geq4$ ($Q\geq16$), a surplus margin (half-gap) in the LP comparison guarantees the inequality; the argument leverages monotonicity in both code length and alphabet size. For qutrits ($q=3$, $V_t(n)$0), the surplus vanishes, requiring a refined spectral analysis using a quadratic-filtered Lloyd square and explicit algebraic positivity arguments (Stein-tangent method). The binary endpoint ($V_t(n)$1, $V_t(n)$2) is controlled separately by a distinct node-edge geometry.

Key Theorems

• Theorem 1: Every exact quantum subspace code $V_t(n)$3 with $V_t(n)$4 satisfies the Hamming bound for $V_t(n)$5.
• Theorem 2: For $V_t(n)$6, the LP comparison produces a uniform half-gap in the critical range $V_t(n)$7 for all $V_t(n)$8.
• Theorem 3: For qutrits ($V_t(n)$9), a quadratic filter and positivity certificate control the critical bridge window (n ~ $q^n$0) without weakening the bound.
• Corollary 4: Combined with an independent binary finite-length theorem (Zhang et al., 14 Jun 2026), the quantum Hamming bound holds for all $q^n$1.

Numerical Results and Explicit Checks

The proof includes strong numerical bounds: for high alphabets, the comparison yields a surplus margin ($q^n$2) rather than just the required threshold. The qutrit bridge region is handled by explicit coefficient-ledgers and Sturm isolation, guaranteeing positivity on all relevant polynomial expressions. Every case is checked algebraically with exact coefficients and explicit sign rules.

Implications for Quantum Error Correction Theory

The result confirms that degeneracy—the phenomenon wherein distinct physical errors induce equivalent actions on a code—cannot, even in fully degenerate or nonadditive settings, create enough overlap to violate the quantum Hamming bound. This generality removes any dependence on code family assumptions (stabilizer, additive, pure), and proves that the sphere-packing intuition remains valid even in the presence of degeneracy, provided exact error correction is required.

On the practical side, the finite-length bound provides definitive limits for code design in all local dimensions, ensuring robust guarantees for quantum memory and communication systems using qudits. Theoretically, the proof techniques bring new mathematical tools to the field—especially the two-center intersection normalized by the Lloyd response and quadratic filters for boundary cases—bridging coding theory, combinatorial geometry, and spectral analysis.

Future Directions

The uniformity of the bound across all local dimensions suggests that similar normalized LP approaches could be fruitfully applied to other quantum code bounds, such as the quantum Johnson and Singleton bounds, or to approximate/impure error correction. The algebraic certificate and positivity arguments developed for the qutrit bridge may inspire further investigation of spectral and polynomial methods in coding theory. The explicit separation of regimes (binary, qutrit, high alphabet) highlights sharp transitions in quantum code geometry, which may be relevant for both theory and device-level implementations.

Conclusion

This paper conclusively proves the quantum Hamming bound in finite length for exact codes of arbitrary local dimension, encompassing fully degenerate and nonadditive structures. The result is universal within the sphere-packing paradigm and offers strong numerical and algebraic confirmation, settling a long-standing open question in quantum error correction theory (2606.22538).