The Bacon-Shor code is a quantum error correcting subsystem code composed of weight 2 check operators that admits a single logical qubit, and has distance $d$ on a $d \times d$ square lattice. We show that when viewed as a Floquet code, by choosing an appropriate measurement schedule of the check operators, it can additionally host several dynamical logical qubits. Specifically, we identify a period 4 measurement schedule of the check operators that preserves logical information between the instantaneous stabilizer groups. Such a schedule measures not only the usual stabilizers of the Bacon-Shor code, but also additional stabilizers that protect the dynamical logical qubits against errors. We show that the code distance of these Floquet-Bacon-Shor codes scales as $\Theta(d/\sqrt{k})$ on a $d \times d$ lattice with $k$ dynamical logical qubits, along with the logical qubit of the parent subsystem code. Moreover, several errors are shown to be self-corrected purely by the measurement schedule itself.
Submitted 05 Mar 2024 to Quantum Physics [quant-ph] Published 07 Mar 2024
Dear authors!
Thanks for posting this interesting paper. I was wondering if you could elaborate on a question I got while reading through. Additionally, I have a minor comment that might be interesting to you.
The question is on the phenomenon you call "self-correcting errors" in your paper. Can you elaborate in which sense the automatic correction of certain errors by the measurement schedule is different to what happens in other measurement-based QEC protocols, e.g. the usual Bacon-Shor Code or topological Floquet codes?
In principle, I understand that this phenomenon can happen in any sequence of projective (Pauli) measurements, also when measuring higher-weight operators (it just happens for higher-weight errors). Is there a particular structure in the code(s) you construct where this happens for more (likely) error combinations and hence constitutes a more relevant process there than in other -- well studied -- error-correcting circuits?
The small comment I have, just for completeness, on the comment "...provides a first example of a Floquet code on a square lattice...". In our paper, https://arxiv.org/abs/2307.11136, we construct a Floquet Code defined on a square lattice. The techniques in that paper can be used to "deform" many other QEC codes/circuits to fit on a square lattice -- or many other connectivities, depending on the connectivity you start with.
Dear Julio, thanks for pointing us to the square lattice construction in your nice paper. We'll be sure to mention it, and modify our statement in a later version.
Re: self-correcting errors, the 2-qubit errors this refers to here are actually check operators of the parent subsystem code, and so do not qualify as errors in the Bacon-Shor code. Here, they do anti-commute with transient stabilizers, and so ought to be considered as errors.
It may be the case that you could always Floquet-ify a subsystem code such that its check operators either always belong in the ISG, or if they flip any stabilizers in some round, they become safely corrected (with or without the measurement schedule) before proliferating into e.g. logical errors. But it's unclear, at least to me, if that would always be the case.
Braneshop, and
the US National Science Foundation.
SciRate
Dear authors!
Thanks for posting this interesting paper. I was wondering if you could elaborate on a question I got while reading through. Additionally, I have a minor comment that might be interesting to you.
The question is on the phenomenon you call "self-correcting errors" in your paper. Can you elaborate in which sense the automatic correction of certain errors by the measurement schedule is different to what happens in other measurement-based QEC protocols, e.g. the usual Bacon-Shor Code or topological Floquet codes?
In principle, I understand that this phenomenon can happen in any sequence of projective (Pauli) measurements, also when measuring higher-weight operators (it just happens for higher-weight errors). Is there a particular structure in the code(s) you construct where this happens for more (likely) error combinations and hence constitutes a more relevant process there than in other -- well studied -- error-correcting circuits?
The small comment I have, just for completeness, on the comment "...provides a first example of a Floquet code on a square lattice...". In our paper, https://arxiv.org/abs/2307.11136, we construct a Floquet Code defined on a square lattice. The techniques in that paper can be used to "deform" many other QEC codes/circuits to fit on a square lattice -- or many other connectivities, depending on the connectivity you start with.
Dear Julio, thanks for pointing us to the square lattice construction in your nice paper. We'll be sure to mention it, and modify our statement in a later version.
Re: self-correcting errors, the 2-qubit errors this refers to here are actually check operators of the parent subsystem code, and so do not qualify as errors in the Bacon-Shor code. Here, they do anti-commute with transient stabilizers, and so ought to be considered as errors.
It may be the case that you could always Floquet-ify a subsystem code such that its check operators either always belong in the ISG, or if they flip any stabilizers in some round, they become safely corrected (with or without the measurement schedule) before proliferating into e.g. logical errors. But it's unclear, at least to me, if that would always be the case.
thanks for the response. It cleared out some confusion about the "self-correcting" errors you mention. Thanks!