arXiv · 2610.06036
Materialised symmetries of 2D translationally invariant codes
Abstract
There has been significant recent interest in near-term qLDPC codes as high-performance alternatives to surface and color codes. One such class of codes is 2-dimensional translationally invariant (TI) codes, such as bivariate bicycle codes, which share similar properties to topological codes. Fundamental objects in the study of such codes are the materialised symmetries, which can be used for the construction of matching-based decoders. These decoders generalise the minimum-weight perfect matching decoder for toric codes and provide similar asymptotic performance guarantees that heuristic decoders such as BP and its variants lack. Despite this, the mathematical structure of symmetries along with their properties under translation and restriction to finite-sized lattices has not been well-studied. We describe a decomposition of symmetry spaces of 2D CSS TI codes on infinite lattices into translation-invariant subspaces which allow us to write an explicit basis of symmetries in a plane-wave-like form. We then describe how to adapt this infinite lattice basis so that any corresponding rectangular periodic lattice basis can be obtained by restriction to basis elements compatible with the periodic boundaries. This allows us to describe how the symmetry space varies with different rectangular dimensions. We illustrate via examples that it is often straightforward in practice to determine a basis of symmetries on twisted boundaries as well, without needing to recompute the decomposition. We comment on the applications of this framework to matching-based decoders and provide examples of symmetries of such codes as the gross code.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Andrew Li, Stephen D. Bartlett, Andrew C. Doherty, Campbell K. McLauchlan. 2026-10-05. Materialised symmetries of 2D translationally invariant codes. https://arxiv.org/abs/2610.06036
Cite the original work for its findings. Save a collection to share your selection of sources.