Decoding quantum error correction with Ising model hardware

Joschka Roffe, Stefan Zohren, Dominic Horsman, Nicholas ChancellorarXiv 2019 · DOI · arXiv

Abstract

Fault-tolerant quantum computers require efficient real-time decoding of quantum error correction (QEC) protocols. This work explores using specialized Ising model hardware—such as quantum annealers and coherent Ising machines—to perform this task. By mapping the decoding problem to an Ising Hamiltonian via the Coherent Parity Check (CPC) framework, we demonstrate that Maximum Entropy (MaxEnt) inference can outperform standard Maximum Likelihood (MLE) decoding, shifting error thresholds and improving error suppression.

Introduction: The Real-Time Decoding Bottleneck

For quantum computers to scale, they must implement fault-tolerant protocols that detect and correct errors in real-time. This involves "decoding": interpreting syndrome measurements (clues about errors) to identify and reverse physical errors. In large-scale systems, decoding must occur at speeds exceeding 100 Gbit/sec. Traditional software-based decoders often fail to meet these latency requirements, creating a computational bottleneck that threatens to stall quantum processors.

Specialized Ising machines—hardware designed to solve optimization problems by finding the ground state of a magnetic model—offer a promising alternative. These devices, including quantum annealers (e.g., D-Wave) and classical CMOS-based Ising chips, can act as high-speed co-processors for the decoding task.

Methodology: Mapping QEC to Ising Models

The core of this approach is the translation of quantum codes into a classical graphical model using the Coherent Parity Check (CPC) framework.

The Ising Mapping

The authors derive a mathematical mapping where the search for an error pattern is transformed into finding the low-energy state of an Ising Hamiltonian. This mapping accounts for complex correlations, such as Y-errors (simultaneous bit and phase flips), which are often simplified or ignored in other models.

The graphical Ising model mapping for QEC decoding, illustrating the lattice of interactions between qubits and parity checks.

Decoding Strategies

Two primary strategies are evaluated within this framework:

  1. Maximum Likelihood Estimation (MLE): This seeks the single most probable error pattern, corresponding to the ground state of the Ising model.
  2. Maximum Entropy Inference (MaxEnt): Rather than a single "best guess," MaxEnt utilizes the full Boltzmann distribution of the Ising model to determine the most likely correction for each individual qubit.

Results: Threshold Shifts and Performance

Numerical simulations on the [[5, 1, 3]] and [[9, 3, 3]] quantum codes reveal that MaxEnt decoding consistently outperforms MLE.

  • Error Thresholds: For the [[5, 1, 3]] code, MaxEnt shifted the error threshold from 7.99% (MLE) to 8.46%. Similarly, for the [[9, 3, 3]] code, the threshold improved from 3.01% to 3.36%.
  • Robustness: The advantage of MaxEnt stems from its ability to "vote" across multiple likely error patterns. In cases where many slightly less likely patterns point to a different correction than the single most likely one, MLE fails while MaxEnt succeeds.

Discussion and Future Outlook

Decoding QEC codes represents a "high-value" application for quantum annealers. Because the structure of a specific quantum code is fixed, hardware can be optimized as an Application Specific Integrated Circuit (ASIC), bypassing the embedding bottlenecks that typically plague general-purpose annealing.

Future work will focus on scaling these mappings to larger codes like the Surface Code and addressing hardware-specific noise within the Ising machines themselves. Additionally, the authors propose a hybrid scheme where classical hardware provides a seed for "reverse annealing," allowing quantum hardware to efficiently sample the Boltzmann distribution for MaxEnt decoding.

Summary updated

Abstract

Fault tolerant quantum computers will require efficient co-processors for real-time decoding of their adopted quantum error correction protocols. In this work we examine the possibility of using specialised Ising model hardware to perform this decoding task. Examples of Ising model hardware include quantum annealers such as those produced by D-Wave Systems Inc., as well as classical devices such as those produced by Hitatchi and Fujitsu and optical devices known as coherent Ising machines. We use the coherent parity check (CPC) framework to derive an Ising model mapping of the quantum error correction decoding problem for an uncorrelated quantum error model. A specific advantage of our Ising model mapping is that it is compatible with maximum entropy inference techniques which can outperform maximum likelihood decoding in some circumstances. We use numerical calculations within our framework to demonstrate that maximum entropy decoding can not only lead to improved error suppression, but can also shift threshold values for simple codes. As a high value problem for which a small advantage can lead to major gains, we argue that decoding quantum codes is an ideal use case for quantum annealers. In addition, the structure of quantum error correction codes allows application specific integrated circuit (ASIC) annealing hardware to reduce embedding costs, a major bottleneck in quantum annealing. Finally, we also propose a way in which a quantum annealer could be optimally used as part of a hybrid quantum-classical decoding scheme.
BibTeX
@misc{roffe2019decoding,
  author = {Joschka Roffe and Stefan Zohren and Dominic Horsman and Nicholas Chancellor},
  title = {{Decoding quantum error correction with Ising model hardware}},
  year = {2019},
  eprint = {1903.10254},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1903.10254},
}
Figure