Hybrid noise protection of logical qubits for universal quantum computation

TL;DR

This paper proposes a hybrid noise protection scheme for logical qubits in quantum computing that combines decoherence-free subspaces and dynamical decoupling, offering a resource-efficient alternative to traditional error correction.

Contribution

It introduces a universal quantum computation model using a two-qubit decoherence-free subspace and DD pulses, avoiding large overheads of standard error correction codes.

Findings

Effective suppression of collective noise using DFS

Numerical evidence supports hybrid protection strategy

Experimental feasibility demonstrated with superconducting qubits

Abstract

Quantum computers now show the promise of surpassing any possible classical machine. However, errors limit this ability and current machines do not have the ability to implement error correcting codes due to the limited number of qubits and limited control. Therefore, dynamical decoupling (DD) and encodings that limit noise with fewer qubits are more promising. For these reasons, we put forth a model of universal quantum computation that has many advantages over strategies that require a large overhead such as the standard quantum error correcting codes. First, we separate collective noise from individual noises on physical qubits and use a decoherence-free subspace (DFS) that uses just two qubits for its encoding to eliminate collective noise. Second, our bath model is very general as it uses a spin-boson type bath but without any Markovian assumption. Third, we are able to either use a steady global magnetic field or to devise a set of DD pulses that remove much of the remaining noise and commute with the logical operations on the encoded qubit. This allows removal of noise while implementing gate operations. Numerical support is given for this hybrid protection strategy which provides an efficient approach to deal with the decoherence problems in quantum computation and is experimentally viable for several current quantum computing systems. This is emphasized by a recent experiment on superconducting qubits which shows promise for increasing the number of gates that can be implemented reliably with some realistic parameter assumptions.