A Practical and Scalable Decoder for Topological Quantum Error Correction with Digital Annealer

TL;DR

This paper introduces a scalable quantum error correction decoder using Fujitsu Digital Annealer, demonstrating improved performance and scalability over traditional methods, with thresholds close to established algorithms.

Contribution

It presents a novel application of Digital Annealer for quantum error correction decoding, showing enhanced scalability and performance for surface codes.

Findings

Lower polynomial scaling compared to SA and MWPM decoders.

Advantages over Union-Find decoder in hardware implementation.

Threshold error rate close to MWPM decoder, between 9.4% and 9.8%.

Abstract

Quantum error correction is one of the most important milestones for realization of large-scale quantum computation. To achieve this, it is essential not only to integrate a large number of qubits with high fidelity, but also to build a scalable classical system that can perform error correction. Here, we propose an efficient and scalable decoder for quantum error correction using Fujitsu Digital Annealer (DA). Specifically, the error correction problem of stabilizer codes is mapped into an Ising-type optimization problem, so-called quadratic unconstrained binary optimization (QUBO) problem, which is solved by DA. In particular, we implement the proposed DA decoder for the surface code and perform detailed numerical experiments for various code distances to see its performance and scalability. We observe that computational scaling for the DA decoder has a lower order of polynomial than the decoding methods using simulated annealing (SA) and minimum-weight perfect matching (MWPM) algorithm under all tested conditions. It is also shown that the DA decoder has advantages over the Union-Find (UF) decoder from a variety of perspectives including hardware implementation. Furthermore, the threshold behavior of the logical error probability for the DA decoder is analyzed and the resultant threshold lies between 9.4% and 9.8%, which is very close to that obtained by the MWPM decoder. This result clearly shows the high potential of the DA decoder for quantum error correction.