Achievable rates for concatenated square Gottesman-Kitaev-Preskill codes
Mahadevan Subramanian, Guo Zheng, Liang Jiang · 2025
Read Achievable rates for concatenated square Gottesman-Kitaev-Preskill codes by Mahadevan Subramanian, Guo Zheng, Liang Jiang (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-2505-10499.txt.
Introduction
Mahadevan Subramanian, Guo Zheng, and Liang Jiang present a 2025 theory paper on quantum error correction. The codes they study, named for Gottesman, Kitaev, and Preskill, store quantum data in a vibrating mode. These codes are known to hit the best possible rates under two kinds of noise, but only at scattered noise levels with the classic design.
The authors join the codes with a second layer of coding to fill the gaps. For one noise type, they pair the code with a quantum polar code and a matching decoder, building the rate step by step. For a lossy channel, they prove by a random coding argument that top rates are reachable at every noise level. The results give fresh ways to build good lattices for these codes.
Catalog record
| CREATOR | Mahadevan Subramanian, Guo Zheng, Liang Jiang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 138 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-10499
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Achievable rates for concatenated square Gottesman-Kitaev-Preskill codes Preprint: APS/123-QED Mahadevan Subramanian, Guo Zheng, Liang Jiang Abstract The Gottesman-Kitaev-Preskill (GKP) codes are known to achieve optimal rates under displacement noise and pure loss channels, which establishes theoretical foundations for its optimality. However, such optimal rates are only known to be achieved at a discrete set of noise strength with the current self-dual symplectic lattice construction. In this work, we develop a new coding strategy using concatenated continuous variable – discrete variable encodings to go beyond past results and establish GKP’s optimal rate over all noise strengths. In particular, for displacement noise, the rate is obtained through a constructive approach by concatenating GKP codes with a quantum polar code and analog decoding. For pure loss channel, we prove the existence of capacity-achieving GKP codes through a random coding approach. These results highlight the capability of concatenation-based GKP codes and provides new methods for constructing good GKP lattices. I Introduction The goal of reliably communicating quantum information through noisy channels lies at the heart of proposals for quantum networks Simon 2017 ; Leung et al. 2010 ; Hayashi et al. 2007 ; Gandotra et al. 2025 ; Kimble 2008 as well as quantum transduction Wang et al. 2022 . A key aspect of this is understanding what rates of information transmission can be considered achievable Barnum et al. 1998 , which has laid the foundation of the study of quantum capacities of noisy channels Lloyd 1997 ; Schumacher and Nielsen 1996 . Showing a rate to be achievable offers a direct lower bound to the capacity of the noise channel. However, explicitly showing achievability of rates requires proving existence of a sequence of encoding and decoding schemes that are asymptotically reliable for that particular rate. Due to the non-trivial nature of this task, the coherent information of a channel is often used to find achievable rates. The coherent information is the quantum analogue to mutual information and offers a lower bound to the quantum capacity Schumacher and Nielsen 1996 . For degradable channels such as the bosonic pure-loss and amplification channels, the maximal one-shot coherent information equals the capacity. Upper and lower bounds on the capacity for channels such as the Gaussian displacement noise Holevo and Werner 1999 ; Harrington and Preskill 2001 , loss-dephasing channel Leviant et al. 2022 and thermal loss channels Noh et al. 2019 have been found while the exact capacity remains unknown. Gottesman-Kitaev-Preskill (GKP) codes Gottesman et al. 2001a are bosonic error correcting codes that encode a finite dimensional Hilbert space into bosonic modes. These codes are obtained by using a stabilizer group generated by optical phase space displacements, hence are deeply connected to symplectically integral lattices Royer et al. 2022 ; Conrad et al. 2022 . This connection has been used to show that GKP codes can achieve the coherent information of the Gaussian displacement noise channel Harrington and Preskill 2001 as well as the capacity of the pure-loss and amplification channels Zheng et al. 2024 . However, these results have two major limitations. The rate R (in qubit per mode) these codes achieve are restricted to have 2^{R} be an integer, and these are existence based results which do not provide constructive ways to obtain rate achieving codes. As a result, the optimal rates are only achieved for a discrete set of noise strengths. All known asymptotic GKP coding rate results Harrington and Preskill 2001 ; Zheng et al. 2024 make use of the family of lattices obtained by rescaling self-dual (unimodular) lattices which as a corollary of the Buser-Sarnak theorem Buser and Sarnak 1994 , will always contain a good spherical packing lattice. This good spherical packing can be used to prove achievability for certain rates. However, to encode information self-dual lattices must be scaled by a square root of an integer (ensuring that it remains valid for encoding a GKP code), which forces the number of logical dimensions per mode to be an integer. Searching for the lattices that satisfy these properties is a hard task Harrington 2004 and the current knowledge of optimal sphere-packing lattices is limited to only dimensions of 24 Cohn et al. 2022 . While the capacity achieving lattices need not be truly optimal in regard to sphere-packing, this does highlight the difficulty one faces in trying to construct such lattices. In this work, we develop a different approach to overcome this limitation, enabling the achievement of favorable rates through more explicit construction methods. We focus on GKP codes that are obtained by taking N single mode square GKP codes, each encoding a qudit of d levels and then defining an outer code for concatenating as a N qudit stabilizer code encoding K qudits which we denote by [[N,K]]_{d} . For the Gaussian displacement noise channel, we first find that the effective noise channel for a single-mode GKP square qudit is a classical mixture of Pauli noise where knowledge of the syndrome gives us knowledge of what the Pauli noise looks like. We formalize this by finding an achievable rate for this noise channel with the use of quantum polar codes Renes et al. 2012 that take advantage of the analog information offered by this syndrome. The usefulness of GKP analog information has been well studied Raveendran et al. 2022 ; Fukui et al. 2017 , to which we show a rigorous way to quantify the advantage analog information offers. We find that this newly defined achievable rate approaches the coherent information of the Gaussian displacement noise channel at all values of noise strength \sigma . Crucially, these GKP codes can be explicitly constructed, and also have efficient ( \mathcal{O}(N\log N) circuit depth) encoding and decoding procedures due to the outer concatenation with the quantum polar code. We further support this with a numerical study of the performance of these polar codes for increasing numbers of modes under the Gaussian displacement noise model. We also show that concatenated GKP codes achieve the capacity of pure-loss, and hence by extension also the capacity of the amplification channel Zheng et al. 2024 . We use a construction of stabilizer codes for prime d dimensional qudits mapped from self-orthogonal codes in the Galois field GF(d^{2}) from Ashikhmin and Knill 2001 . By averaging over the set of GKP codes obtained in this manner, we prove existence of a capacity achieving sequence of codes as d\to\infty , in essence recreating the behavior of a good spherical packing lattice. These results highlight the importance of the family of codes obtained by simply concatenating qudit codes to square GKP codes, as well as their capability in exceeding previously known achievable rates for GKP codes. This paper is structured as follows. In section II.1 , we introduce the square GKP code, following which we provide a short introduction to achievable rates for communication of quantum information in section II.2 . Our results begin with section III.1 where show how the effective logical noise channel of a corrected square GKP qudit is a classical mixture of Pauli noise with an auxiliary output containing syndrome information. We then show in section III.2 how the coherent information of this logical noise channel which we refer to as I^{\mathrm{sq}}_{d,\mathrm{analog}} approaches the