Growing Sparse Quantum Codes from a Seed

TL;DR

This paper introduces a method to systematically construct quantum LDPC and subsystem codes by conjoining simple repetition codes, enabling the creation of high-quality quantum codes with bounded degree Tanner graphs.

Contribution

It presents an efficient iterative algorithm for building sparse quantum codes with near-optimal distance properties using a generalized concatenation called conjoining.

Findings

Quantum LDPC codes can be constructed by conjoining repetition codes.

The conjoining of two-qubit repetition codes can generate any CSS code.

The method achieves asymptotic saturation of the code distance bound.

Abstract

It is generally unclear whether smaller codes can be "concatenated" to systematically create quantum LDPC codes or their sparse subsystem code cousins where the degree of the Tanner graph remains bounded while increasing the code distance. In this work, we use a slight generalization of concatenation called conjoining introduced by the quantum lego formalism. We show that by conjoining only quantum repetition codes, one can construct quantum LDPC codes. More generally, we provide an efficient iterative algorithm for constructing sparse subsystem codes with a distance guarantee that asymptotically saturates $kd^2=O(n)$ in the worst case. Furthermore, we show that the conjoining of even just two-qubit quantum bit-flip and phase-flip repetition codes is quite powerful as they can create any CSS code. Therefore, more creative combinations of these basic code blocks will be sufficient for generating good quantum codes, including good quantum LDPC codes.