| name | spatially-coupled-quantum-codes |
| description | Spatial coupling methodology for quantum LDPC/CSS codes. Proves that belief-propagation decoding on spatially coupled CSS codes (MN/HA-type) achieves the quantum erasure hashing bound. Uses coupled-vector potential method and density evolution analysis to show BP threshold equals MAP threshold. |
Spatially Coupled Quantum Codes
Description
Spatial coupling is a technique from classical coding theory adapted to quantum error correction. By spatially coupling CSS codes (Calderbank-Shor-Steane), belief-propagation (BP) decoding achieves the maximum-a-posteriori (MAP) performance — specifically the quantum erasure hashing bound — despite BP being a suboptimal decoder. This methodology provides a DE-level proof for seeded BP decoding on finite-degree factor graphs.
Activation Keywords
- spatially coupled quantum code
- spatial coupling CSS
- quantum erasure hashing bound
- seeded belief propagation
- MN/HA CSS code
- coupled vector potential
- density evolution quantum
- 空间耦合量子码
- 量子删除信道哈希界
Core Concepts
Spatial Coupling
- A technique that couples multiple instances of a base code along a chain
- Enables BP decoding to achieve MAP threshold (threshold saturation)
- Originally discovered for classical LDPC codes; extends to quantum CSS codes
CSS Codes from MN/HA Ensembles
- MacKay-Neal/Hsu-Anastasopoulos punctured sparse ensembles
- Achieve capacity under MAP decoding
- Spatial coupling + seeded BP = achieves hashing bound under BP
Quantum Erasure Channel & Hashing Bound
- On the quantum erasure channel, erased qubits are replaced by maximally mixed states
- Hashing bound: R = 1 - 2p (for erasure probability p)
- Achieving the hashing bound means optimal performance for the given code rate
Density Evolution (DE) Analysis
- Tracks message distributions through BP iterations
- Five-message DE recursion for CSS erasure decoding
- Decomposes into Z-side and X-side constituent systems
Methodology
Step 1: CSS Ensemble Definition
Define the CSS code ensemble:
- Specify sparse punctured parity-check matrices for X and Z checks
- Define corresponding dense parity-check matrices
- Set finite Z-side degree, X-side degree, and check degrees
Step 2: Erasure Model Setup
On an erased coordinate:
- Two binary Pauli components remain unresolved
- Erased qubit represented by four Pauli possibilities (I, X, Y, Z)
- Map to hard-erasure CSS decoding problem
Step 3: Density Evolution Derivation
Derive the DE recursion:
- Write five-message DE recursion for the uncoupled system
- Decompose into Z-side and X-side constituent systems
- Define two constituent potentials (one per side)
Step 4: Coupled-Vector Potential Analysis
Apply the coupled-vector potential method:
- Apply to Z-side constituent → Z-side BP threshold
- Apply to X-side constituent → X-side BP threshold
- Combined threshold = min(Z-side degree ratio, X-side complementary degree ratio)
Step 5: Hashing Bound Achievement
For X/Z equal-rate specialization:
- Z-side and X-side constituent design rates are equal
- BP threshold = hashing-bound channel parameter determined by design rate
- Proves seeded BP achieves hashing bound on finite-degree factor graphs
Workflow
Pattern 1: Spatially Coupled CSS Code Design
For constructing spatially coupled CSS codes:
- Choose base MN/HA ensemble parameters (Z-degree, X-degree, check degree)
- Define spatial coupling window and chain length
- Design seeded BP decoder with finite-degree factor graphs
- Verify DE threshold equals hashing bound for equal-rate case
Pattern 2: Erasure Channel Performance Analysis
For analyzing erasure channel performance:
- Model erasure probability distribution
- Run DE analysis for the coupled system
- Compare BP threshold vs MAP threshold
- Verify threshold saturation (BP → MAP)
Error Handling
Finite-Length Effects
DE analysis assumes infinite block length. For finite-length codes:
- Use BP concentration results for finite-block analysis
- Consider block-error convergence behavior
- Design finite-code realization of ideal DE seed
Unequal X/Z Rates
For unequal X/Z design rates:
- BP threshold = min(Z-side degree ratio, X-side complementary degree ratio)
- Not equal to hashing bound; requires different analysis
Resources
- Paper: arXiv:2606.32001
- Related:
quantum-error-correction-methods (umbrella), loss-biased-qec (bias-tailored QEC), coset-based-qldpc-codes (CSS code construction)