Optimized Noise Suppression for Quantum Circuits

TL;DR

This paper presents an efficient method for characterizing and mitigating crosstalk noise in quantum circuits, improving quantum algorithm performance by integrating noise data into qubit routing.

Contribution

It introduces a simplified crosstalk measurement, an optimized experiment scheduling, and a noise-aware qubit routing algorithm with proven convex hull properties, enhancing quantum circuit fidelity.

Findings

Characterized crosstalk noise on 127-qubit chips.

Improved Quantum Approximate Optimization Algorithm performance by up to 10%.

Demonstrated the effectiveness of noise-aware qubit routing.

Abstract

Quantum computation promises to advance a wide range of computational tasks. However, current quantum hardware suffers from noise and is too small for error correction. Thus, accurately utilizing noisy quantum computers strongly relies on noise characterization, mitigation, and suppression. Crucially, these methods must also be efficient in terms of their classical and quantum overhead. Here, we efficiently characterize and mitigate crosstalk noise, which is a severe error source in, e.g., cross-resonance based superconducting quantum processors. For crosstalk characterization, we develop a simplified measurement experiment. Furthermore, we analyze the problem of optimal experiment scheduling and solve it for common hardware architectures. After characterization, we mitigate noise in quantum circuits by a noise-aware qubit routing algorithm. Our integer programming algorithm extends previous work on optimized qubit routing by swap insertion. We incorporate the measured crosstalk errors in addition to other, more easily accessible noise data in the objective function. Furthermore, we strengthen the underlying integer linear model by proving a convex hull result about an associated class of polytopes, which has applications beyond this work. We evaluate the proposed method by characterizing crosstalk noise for two chips with up to 127 qubits and leverage the resulting data to improve the approximation ratio of the Quantum Approximate Optimization Algorithm by up to 10 % compared to other established noise-aware routing methods. Our work clearly demonstrates the gains of including noise data when mapping abstract quantum circuits to hardware native ones.