A Class of Cyclic Quantum Codes
Matthew B. Hastings · 2025
Read A Class of Cyclic Quantum Codes by Matthew B. Hastings (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-2509-06865.txt.
Introduction
Quantum error-correcting codes built around easy state preparation are introduced by Matthew B. Hastings in a solo 2025 quantum information paper. Most code families are designed for simple stabilizers, the checks that detect errors. Hastings takes the opposite organizing principle: simplicity of preparing a code state, at least when no noise is present.
Certain known codes turn out to belong to this cyclic class, including a family of rotated two-dimensional toric codes. Other small examples were found by computer search, showing the class is not just a redescription of one known family. The paper closes by discussing how to prepare these codes fault-tolerantly, so that the preparation process itself does not introduce errors faster than the code can handle. Preparation-first design could matter for near-term devices, where state preparation is a practical bottleneck.
Catalog record
| CREATOR | Matthew B. Hastings |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-06865
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Class of Cyclic Quantum Codes Matthew B. Hastings Abstract We introduce a class of cyclic quantum codes, basing the construction not on the simplicity of the stabilizers, but rather on the simplicity of preparation of a code state (at least in the absence of noise). We show how certain known codes, such as a certain family of rotated two-dimensional toric codes, fall into this class, and we also give certain other examples at small sizes found by computer search. We finally discuss fault tolerant preparation of these codes. We introduce a class of cyclic quantum codes defined using a bipartite cluster state. After a Hadamard transformation on certain qubits, these are CSS quantum codes. We term these codes BCC codes (“bipartite cyclic cluster” codes). Generally, for quantum stabilizer codes, there is some interest in finding “simpler” codes. This has often been interpreted as wanting the stabilizers to be low weight, as this simplifies the circuits to measure the syndrome. Here, we consider a different notion of a simpler code, asking for simplified circuits to prepare a code state initialized with a particular value of logicals. We give a general definition of these codes in Section I and specific properties, followed by showing that certain known codes fall into this class in Section II , and giving some small examples found by computer search in Section III . For the majority of this, we will consider preparation circuits in the absence of noise, only considering fault tolerant preparation in Section IV . The family of codes considered here has some advantages. They may provide simpler preparation circuits. They are rich enough to exhibit reasonable distance, and yet they are structured enough that they are amenable to computer search. Their structure may be useful in implementations in neutral atom systems where the atoms are moved near each other before implementing gates between them, as this structure may simplify the moves required. Some of the BCC codes are a special case of bicycle codes in Kovalev and Pryadko 2013 ; the interest here is the emphasis on the simpler preparation circuit, meaning that some codes with relatively high weight stabilizers (and so being relatively complicated as bicycle codes) are simple from this point of view. I General Definition In general, we will consider [[n,k,d]] codes where n=kn_{0} for some given n_{0} . We label the qubits by a pair (m,j) where m is defined mod n_{0} and where 1\leq j\leq k . For a given code, define a basis of code states as follows; these basis states are those which can be prepared by, for each j , initializing all qubits (m,j) in either the + or the - state, and then applying a certain unitary U , so that there are 2^{k} basis states. We require that the unitary U be a product of CZ gates obeying certain properties below. Because the state is created by initializing qubits to + (or - ) and applying a product of CZ gates, the resulting state is a cluster state Briegel and Raussendorf 2001 (or possibly a cluster state up to applying Pauli Z to certain qubits which are initialized in the - state). Such a cluster state can be described by a graph G : vertices correspond to qubits and there is an edge between vertices if a CZ gates acts on those two qubits (we assume that for each pair of qubits, there are either zero or one CZ gates acting on that pair). To state the properties of U in terms of this graph G , this graph should be bipartite and have a certain cyclic invariance. Formally: to define bipartiteness, we define some set {\cal A}\subset\{1,\ldots,k\} and we define the set {\cal B} to be \{1,\ldots,k\}\setminus{\cal A} . We define two sets of qubits A,B , where A is the set of qubits (m,j) with j\in{\cal A} and B is the set of qubits (m,j) with j\in{\cal B} . Then we require that (1): each CZ gate acts on one qubit in A and one qubit in B , and (2): U is cyclically invariant, in that if a gate acts on qubits (m,j) and (m^{\prime},j^{\prime}) then for each l , there is a gate acting on (m+l,j) and (m^{\prime}+l,j^{\prime}) . Thus, the parameters of the code are n_{0},k , and the given graph G . This code indeed has the given number of logical qubits k as for each j\in\{1,\ldots,k\} we may choose to either initialize in + or in - . By applying a Hadamard on qubits in B to the resulting code state, the code is a CSS quantum code. The resulting code can be described by initializing qubits in A in + or - for each j , initializing qubits in B in 0 or 1 for each j , calling the resulting state the “initial state”,and then applying CNOT gates with source in A and target in B , calling the resulting state the “code state”. F or the rest of this paper, we will use this CSS quantum code. Note that the initial state is a repetition code in the +,- basis for qubits in A and in the 0,1 basis for qubits in B . We may compute the stabilizers of the code as follows. For each j\in{\cal A} , for each m , the operator X_{m,j}X_{m+1,j} is a stabilizer of the initial state. Hence, UX_{m_{j}}X_{m+1,j}U^{\dagger} is a stabilizer of the code state. This operator is X_{m,j}X_{m+1,j} times a product of X over qubits in B whose corresponding vertices share an odd number of edges with (m,j) or (m+1,j) . Similarly, for each j\in{\cal B} , the operators Z_{m,j}Z_{m+1,j} is a stabilizer of the initial state and therefore a stabilizer of the code state is Z_{m,j}Z_{m+1,j} times a product of Z over qubits in B whose corresponding vertices share an odd number of edges with (m,j) or (m+1,j) . From the stabilizers of the code, we may see that these BCC codes are a special case of generalized bicycle codes in Kovalev and Pryadko 2013 , with a choice of A having weight two. The particular interest here is the simple class of circuit giving these codes. We may easily vertify from the redundancy of the stabilizers that the number of the logical qubits indeed equals k : the number of stabilizers given equals the number of qubits while the product of these stabilizers over m for a given j is equal to the identity, so there are j redundancies among the stabilizers (and no other redundancies as may easily be seen). Note of course that in general UX_{m,j}X_{m+l,j}U^{\dagger} for j\in{\cal A} is a stabilizer of the code, for any l , and similarly for j\in{\cal B} with X replaced with Z . This will be useful later. F or the rest of this paper, we focus on the case k=2 . We will assume {\cal A} is the set \{1\} while {\cal B} is the set \{2\} . In this case, one interesting transversal (up to permutation of qubits) logical operation always exists. If we interchange qubits (m,1) and (-m,2) , then the corresponding interchange of vertices gives an automorphism of G