Some aspects of noise in binary classification with quantum circuits

TL;DR

This paper investigates how a specific single-qubit noise model and data corruption affect quantum circuit-based binary classification, revealing that noise can sometimes serve as a beneficial regularizer.

Contribution

It provides a formal analysis of noise effects in quantum binary classification, highlighting that noise on a qubit affects only that qubit and can act as a regularizer during training.

Findings

Measurement noise affects only the targeted qubit even with entanglement

Noise in training data can serve as a regularizer

Potential benefits of noise in quantum machine learning

Abstract

We formally study the effects of a restricted single-qubit noise model inspired by real quantum hardware, and corruption in quantum training data, on the performance of binary classification using quantum circuits. We find that, under the assumptions made in our noise model, that the measurement of a qubit is affected only by the noises on that qubit even in the presence of entanglement. Furthermore, when fitting a binary classifier using a quantum dataset for training, we show that noise in the data can work as a regularizer, implying potential benefits from the noise in certain cases for machine learning problems.