Uncertainty of Feed Forward Neural Networks Recognizing Quantum Contextuality

TL;DR

This paper enhances neural network performance in quantum contextuality recognition by integrating uncertainty estimation via Bayesian neural networks, enabling more reliable predictions even with biased training data.

Contribution

It introduces the use of Bayesian neural networks for quantum contextuality recognition, emphasizing their ability to estimate uncertainty and improve classification reliability.

Findings

BNNs provide reliable uncertainty estimates in quantum tasks.

Uncertainty correlates with misclassification risk.

BNNs outperform traditional methods in biased data scenarios.

Abstract

The usual figure of merit characterizing the performance of neural networks applied to problems in the quantum domain is their accuracy, being the probability of a correct answer on a previously unseen input. Here we append this parameter with the uncertainty of the prediction, characterizing the degree of confidence in the answer. A powerful technique for estimating both the accuracy and the uncertainty is provided by Bayesian neural networks (BNNs). We first give simple illustrative examples of advantages brought forward by BNNs, out of which we wish to highlight their ability of reliable uncertainty estimation even after training with biased data sets. Then we apply BNNs to the problem of recognition of quantum contextuality which shows that the uncertainty itself is an independent parameter identifying the chance of misclassification of contextuality.