The Role of Graph Theory in Explainable AI (XAI)

Abstract

These days, deep neural networks (DNNs) are used widely in autonomous vehicles, healthcare, military and other mission-critical systems with direct effect on lives of people.  Its use is challenged by the black-box problem of DNN, raising regulatory and ethical concerns like lack of trust. In “Artificial Intelligence (AI)”, “Explainable AI (XAI)” promotes a range of techniques, tools, and models to generate high-quality, intuitive, interpretable, and simple explanations of AI decisions.