XYZ ruby code: Making a case for a three-colored graphical calculus for quantum error correction in spacetime

TL;DR

This paper introduces a tensor network graphical formalism for analyzing Clifford quantum error correction protocols with explicit time dynamics, applied to new Floquet codes called XYZ ruby codes, showing promising error thresholds.

Contribution

It develops a novel tensor network-based graphical framework for understanding and designing dynamic quantum error correction codes, including Pauli flows and decoding methods.

Findings

Introduces a tensor network formalism for Clifford circuits with measurements.

Develops Pauli flows for graphical error correction analysis.

Achieves competitive noise thresholds of 0.18% on torus simulations.

Abstract

Analyzing and developing new quantum error-correcting schemes is one of the most prominent tasks in quantum computing research. In such efforts, introducing time dynamics explicitly in both analysis and design of error-correcting protocols constitutes an important cornerstone. In this work, we present a graphical formalism based on tensor networks to capture the logical action and error-correcting capabilities of any Clifford circuit with Pauli measurements. We showcase the formalism on new Floquet codes derived from topological subsystem codes, which we call XYZ ruby codes. Based on the projective symmetries of the building blocks of the tensor network we develop a framework of Pauli flows. Pauli flows allow for a graphical understanding of all quantities entering an error correction analysis of a circuit, including different types of QEC experiments, such as memory and stability experiments. We lay out how to derive a well-defined decoding problem from the tensor network representation of a protocol and its Pauli flows alone, independent of any stabilizer code or fixed circuit. Importantly, this framework applies to all Clifford protocols and encompasses both measurement- and circuit-based approaches to fault tolerance. We apply our method to our new family of dynamical codes which are in the same topological phase as the 2+1d color code, making them a promising candidate for low-overhead logical gates. In contrast to its static counterpart, the dynamical protocol applies a Z3 automorphism to the logical Pauli group every three timesteps. We highlight some of its topological properties and comment on the anyon physics behind a planar layout. Lastly, we benchmark the performance of the XYZ ruby code on a torus by performing both memory and stability experiments and find competitive circuit-level noise thresholds of 0.18%, comparable with other Floquet codes and 2+1d color codes.