Sequential decoding of the XYZ 2 hexagonal stabilizer code
Basudha Srivastava, Yinzi Xiao, Anton Frisk Kockum, Ben Criger, Mats Granath · 2025
Read Sequential decoding of the XYZ 2 hexagonal stabilizer code by Basudha Srivastava, Yinzi Xiao, Anton Frisk Kockum, Ben Criger, Mats Granath (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-03691.txt.
Introduction
Basudha Srivastava and four co-authors present a 2025 study of decoding a honeycomb quantum error-correcting code in two sequential steps. The XYZ2 code can be read as two nested codes: a tiny phase-flip check inside a surface code variant. That structure lets a decoder pass error probabilities upward, level by level, instead of solving everything at once. The team tests maximum-likelihood and cheaper matching-based decoders under several noise shapes.
For standard noise, the matching decoder reached an error threshold of 18.3 percent, near optimal. Under phase-heavy noise, a belief-matching decoder hit 24.1 percent, against 28.6 for the costly optimal decoder. Measurement errors lowered thresholds to a few percent. The authors say the code competes with designs built for biased noise, and the nesting trick is reusable.
Catalog record
| CREATOR | Basudha Srivastava, Yinzi Xiao, Anton Frisk Kockum, Ben Criger, Mats Granath |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-03691
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Sequential decoding of the XYZ 2 hexagonal stabilizer code Basudha Srivastava, Yinzi Xiao, Anton Frisk Kockum, Ben Criger, Mats Granath Abstract Quantum error correction requires accurate and efficient decoding to optimally suppress errors in the encoded information. For concatenated codes, where one code is embedded within another, optimal decoding can be achieved using a message-passing algorithm that sends conditional error probabilities from the lower-level code to a higher-level decoder. In this work, we study the XYZ 2 topological stabilizer code, defined on a honeycomb lattice, and use the fact that it can be viewed as a concatenation of a ⟦ 2 , 1 , 1 ⟧ \left\llbracket 2,\,1,\,1\right\rrbracket phase-flip parity check code and the surface code with Y Z Z Y YZZY stabilizers, to decode the syndrome information in two steps. We use this sequential decoding scheme to correct errors on data qubits, as well as measurement errors, under various biased error models using both a maximum-likelihood decoder (MLD) and more efficient matching-based decoders. For depolarizing noise we find that the sequential matching decoder gives a threshold of 18.3 % 18.3\text{\,}\mathrm{\%} , close to optimal, as a consequence of a favorable, effectively biased, error model on the upper-level YZZY code. For phase-biased noise on data qubits, at a bias η = p z p x + p y = 10 \eta=\frac{p_{z}}{p_{x}+p_{y}}=10 , we find that a belief-matching-based decoder reaches thresholds of 24.1 % 24.1\text{\,}\mathrm{\%} , compared to 28.6 % 28.6\text{\,}\mathrm{\%} for the MLD. With measurement errors the thresholds are reduced to 3.4 % 3.4\text{\,}\mathrm{\%} and 4.3 % 4.3\text{\,}\mathrm{\%} , for depolarizing and biased noise respectively, using the belief-matching decoder. This demonstrates that the XYZ 2 code has thresholds that are competitive with other codes tailored to biased noise. The results also showcase two approaches to taking advantage of concatenated codes: 1) tailoring the upper-level code to the effective noise profile of the decoded lower-level code, and 2) making use of an upper-level decoder that can utilize the local information from the lower-level code. I Introduction Quantum computers are expected to have a computational advantage over classical computers for certain applications Shor 1994 ; Grover 1997 ; Jordan ; Montanaro 2016 ; Dalzell et al. 2023 . To achieve these advantages, quantum computers must operate with significantly lower error rates than what can currently be achieved experimentally Fowler et al. 2012 ; Preskill 2018 ; Kivlichan et al. 2020 ; Campbell 2021 . Overcoming the adverse effects of noise requires the use of quantum error-correcting codes, that will come with an overhead in space (number of qubits) and/or time (number of gates) Nielsen and Chuang 2000 ; Devitt et al. 2013 ; Terhal 2015 . One of the most prominent approaches to quantum error correction (QEC) uses topological stabilizer codes Kitaev 2003 which are constructed by tiling a manifold. The number of logical qubits in these codes is determined by the topology of the manifold, while the code distance is determined by the density of the tiling. Topological codes have several favorable properties, such as requiring only local entangling gates, having low-depth syndrome measurement circuits, and high threshold error rates (below which lower logical error rates can be achieved by increasing code distance). Among these, the surface code Fowler et al. 2012 is the most well known and has seen a growing number of experimental demonstrations in recent years Andersen et al. 2020 ; Satzinger et al. 2021 ; Marques et al. 2021 ; Krinner et al. 2022 ; Acharya et al. 2023 ; Acharya et al. 2025 . Another approach to achieving fault tolerance uses concatenated codes — errors at the physical level are suppressed using an error-correcting code, and the resulting less noisy logical qubits are then re-encoded into higher-level codes. This approach offers several advantages, including better architecture when using short- and long-range gates Criger and Terhal 2016 , constant space overhead in the absence of connectivity constraints Yamasaki and Koashi 2024 , and constant-factor reductions in space overhead in planar architectures Gidney et al. 2023 , among others Pattison et al. 2023 ; Yoshida et al. 2024 ; Vasmer and Kubica 2022 ; Goto 2024 ; Li et al. 2023 ; Meister et al. 2024 . Due to the intrinsic noise bias found in various qubit architectures, another line of recent research is the design of QEC codes tailored to such noise models Bonilla Ataides et al. 2021 ; Hastings and Haah 2021 ; Haah and Hastings 2022 ; Wootton 2015 ; Srivastava et al. 2022 ; Huang et al. 2023 ; Wootton 2022 . In this work, we focus on the XYZ 2 code Wootton 2015 ; Wootton 2022 ; Srivastava et al. 2022 , a topological stabilizer code with favorable properties under biased noise. The XYZ 2 code has local stabilizers and a code distance which is linear in the number of physical qubits for pure Z noise. It has previously been shown to have high thresholds and highly suppressed logical failure rates for phase-biased noise, under the assumption of perfect stabilizer measurements Srivastava et al. 2022 . In this work, we use the fact that XYZ 2 code can be described as a concatenation of a \left\llbracket 2,\,1,\,1\right\rrbracket phase-flip error-detecting code embedded into a surface code where each plaquette of four qubits is stabilized by the product of Pauli matrices YZZY (see Fig. 1 ), here referred to as the YZZY code Bonilla Ataides et al. 2021 . The latter is an example of a Clifford-deformed surface code Bonilla Ataides et al. 2021 ; Tuckett et al. 2018 ; Tuckett et al. 2019 ; Tuckett et al. 2020 ; Dua et al. 2024 ; Xiao et al. 2024 , which in itself has favorable properties for phase-biased noise. The concatenated structure allows us to implement sequential decoders for the XYZ 2 code using Poulin’s message-passing technique Poulin 2006 . Figure 1: The XYZ 2 code can be viewed as a concatenation of the YZZY code with each data qubit in the higher-level YZZY code encoded in a lower-level \left\llbracket 2,\,1,\,1\right\rrbracket phase-flip error-detecting code, with a single XX stabilizer. The figure depicts the transformation from an XYZXYZ plaquette stabilizer to a YZZY plaquette stabilizer by replacing each pair of physical qubits (white circles) on a link (orange ellipse) with a logical qubit (orange circle) of the \left\llbracket 2,\,1,\,1\right\rrbracket code. We make use of matching Higgott and Gidney 2022 , belief-matching Criger and Ashraf 2018 ; Higgott et al. 2023 , and matrix-product-state-based Bravyi et al. 2014a ; Tuckett 2020 ; Xiao et al. 2024 decoders for the upper-level (YZZY) code under code-capacity noise (where errors may occur on data qubits, but not measurement results) and phenomenological noise (where errors occur on data qubits and measurement results independently). We find that a standard matching-based decoder that uses the conditional logical error probabilities from the lower-level syndrome measurement performs nearly optimally for code-capacity depolarizing noise. This is a consequence of a resulting biased noise profile on the upper-level YZZY code, for which the latter is well suited. For phase-biased noise on the XYZ 2 code, the noise on the higher-level code is effectively X -biased. Here, pre-processing with belief propagation provides better input probabilities