Gaussian breeding for encoding a qubit in propagating light

TL;DR

This paper introduces Gaussian breeding, a method to encode GKP qubits in propagating light using photon detectors, enabling scalable and high-fidelity quantum information processing for practical quantum computing.

Contribution

The paper presents a novel Gaussian breeding technique for encoding GKP qubits in propagating light, overcoming nonlinear challenges with efficient photon detection.

Findings

GKP qubits above fault-tolerance threshold can be generated with high success probability.

High-fidelity GKP qubits exceeding 0.99 fidelity are achievable.

The method enables systematic creation of quantum superpositions with minimal resources.

Abstract

Practical quantum computing requires robust encoding of logical qubits in physical systems to protect fragile quantum information. Currently, the lack of scalability limits the logical encoding in most physical systems, and thus the high scalability of propagating light can be a game changer for realizing a practical quantum computer. However, propagating light also has a drawback: the difficulty of logical encoding due to weak nonlinearity. Here, we propose Gaussian breeding that encodes arbitrary Gottesman-Kitaev-Preskill (GKP) qubits in propagating light. The key idea is the efficient and iterable generation of quantum superpositions by photon detectors, which is the most widely used nonlinear element in quantum propagating light. This formulation makes it possible to systematically create the desired qubits with minimal resources. Our simulations show that GKP qubits above a fault-tolerant threshold, including ``magic states'', can be generated with a high success probability and with a high fidelity exceeding 0.99. This result fills an important missing piece toward practical quantum computing.