Bayesian Networks can be used for machine learning to predict whether or not an email is spam. While classical Bayesian Networks can operate quickly and provide satisfactory accuracy in spam predictions, Quantum Bayesian Networks consistently exceed classical results and can provide better accuracy along with faster training times.
Spam emails are unsolicited emails with intentions ranging from mere nuisance to significant cybersecurity threats. Types of spam emails include commercial advertisements, phishing, money scams, and antivirus warnings. In order to drastically reduce the number of spam emails one receives, machine learning algorithms can be used. With effective classification programs involving machine learning algorithms such as Naïve Bayes, potential for harm caused by these emails can be minimized.
A Bayesian Belief Network is a type of probabilistic graphical model. A probabilistic graph model is a method of representing a probabilistic model in a graph format. It can be used for tasks such as prediction, diagnostics, and reasoning in a wide array of subjects including prognostics, health monitoring, system diagnostics, and performance evaluation. The Bayesian Network graph representation consists of a directed acyclic graph with nodes representing variables and edges representing the probabilistic dependence between the nodes.
Some issues have arisen from the implementation of classical Bayesian Networks, most notably the high computational expense as a result of a significant amount of nodes involved in a particular network. A method to resolve this is to employ the use of quantum computing. This is because while classical computer bits hold values of either 0 or 1, qubits (quantum bits) have the ability to hold 0, 1, or both in the form of superposition. It has been shown that involving the use of quantum computing can increase performance in many areas including result accuracy and time required to compute. In the case of Bayesian Networks involving spam email classification, Quantum Networks can be shown to have superiority in the aforementioned performance aspects.
Brownlee, J. (2019, September 24). A Gentle Introduction to Bayesian Belief Networks. MachineLearningMastery.com. Retrieved April 2023, from https://machinelearningmastery.com/introduction-to-bayesian-belief-networks/
Borujeni, S. E., Nannapaneni, S., Nguyen, N. H., Behrman, E. C., & Steck, J. E. (2021, April 12). Quantum Circuit representation of Bayesian Networks. arXiv.org. Retrieved April 2023, from https://arxiv.org/abs/2004.14803