Source-linked AI summary
Neural Belief-Propagation Decoders for Quantum Error-Correcting Codes
Ye-Hua Liu, David Poulin
TL;DR
BP is inadequate for quantum LDPC decoding because quantum error degeneracy makes stabilizer-equivalent errors valid even when they differ from the true error pattern. The paper trains neural BP decoders with a degeneracy-aware loss and reports orders-of-magnitude accuracy gains across the tested code families.
Problem
BP is inadequate for quantum LDPC decoding because success permits stabilizer-equivalent errors, unlike exact-pattern matching, creating an error-degeneracy challenge.
Source: pp. 1–3Method
The paper trains neural BP decoders for quantum LDPC codes using a loss function that accounts for error degeneracy.
Source: p. 1Results
Across toric, bicycle, and hypergraph-product codes, training yields orders-of-magnitude decoding-accuracy enhancement over untrained BP.
Source: pp. 1, 4Takeaways & Limitations
Combining machine learning's general framework with quantum-error-correction domain knowledge is promising when neither works well alone.
Source: p. 4The approach does not completely solve LDPC decoding.
Source: p. 1
Abstract
from arXiv · showhide
Belief-propagation (BP) decoders play a vital role in modern coding theory, but they are not suitable to decode quantum error-correcting codes because of a unique quantum feature called error degeneracy. Inspired by an exact mapping between BP and deep neural networks, we train neural BP decoders for quantum low-density parity-check (LDPC) codes with a loss function tailored to error degeneracy. Training substantially improves the performance of BP decoders for all families of codes we tested and may solve the degeneracy problem which plagues the decoding of quantum LDPC codes.