Title:Quantum Bayesian Networks for Machine Learning in Oil-Spill Detection

Abstract:Quantum Machine Learning (QML) has shown promise in diverse applications such as environmental monitoring, healthcare diagnostics, and financial modeling. However, the practical application of QML faces challenges, such as the limited availability of quantum hardware and the complexity of integrating quantum algorithms with classical systems. This paper introduces a novel Bayesian approach using Quantum Bayesian Networks (QBNs) to classify imbalanced datasets, focusing on differentiating ``oil-spill'' from ``non-spill'' classes in satellite-derived data. By employing QBNs, which combine probabilistic reasoning with quantum state preparation, we effectively address the challenge of integrating quantum enhancements with classical machine learning architectures. While the integration improves key performance metrics, it also uncovers areas for refinement, highlighting the need for customized strategies to address specific challenges and optimize outcomes. Our study demonstrates significant advances in detecting and classifying anomalies, contributing to more effective and precise environmental monitoring and management.