Machine Learning Methods as Robust Quantum Noise Estimators

TL;DR

This work addresses the challenge of estimating quantum circuit noise to support production-ready quantum software. It proposes two data representations—gate-count based tabular features and image-like circuit embeddings—and trains a mix of tabular (Linear Regression, HistGradientBoosting, XGBoost) and neural network models (Dense NN, CNN) on a large set of 35,000 random circuits simulated with IBM noise models using Qiskit Aer. The results show that XGBoost and CNN-based approaches can predict the distance between ideal and noisy outputs with low error, and that circuit components such as the total qubit count heavily influence noise under certain backends; removing qubit-count related gates reduces this effect. The methods offer fast inference and demonstrate potential for integration into production-grade quantum code quality tools, enabling scalable assessment of circuit robustness and reliability.

Abstract

Access to quantum computing is steadily increasing each year as the speed advantage of quantum computers solidifies with the growing number of usable qubits. However, the inherent noise encountered when running these systems can lead to measurement inaccuracies, especially pronounced when dealing with large or complex circuits. Achieving a balance between the complexity of circuits and the desired degree of output accuracy is a nontrivial yet necessary task for the creation of production-ready quantum software. In this study, we demonstrate how traditional machine learning (ML) models can estimate quantum noise by analyzing circuit composition. To accomplish this, we train multiple ML models on random quantum circuits, aiming to learn to estimate the discrepancy between ideal and noisy circuit outputs. By employing various noise models from distinct IBM systems, our results illustrate how this approach can accurately predict the robustness of circuits with a low error rate. By providing metrics on the stability of circuits, these techniques can be used to assess the quality and security of quantum code, leading to more reliable quantum products.

Figures (1)

• Figure 1: Schema detailing the process followed to obtain the data and train the different models.