Buxton R. 2022. Graph models and explainable AI. Lower Hutt (NZ): GNS Science. 42 p. (GNS Science report; 2022/17). doi:10.21420/RVE8-DX15.
Abstract
This work presents a brief review of different graph models and then discusses how they can be utilised as an approach for Explainable Artificial Intelligence (XAI). XAI is a branch of computer science that is gaining in popularity, as it addresses the fundamental concept of ‘trust in AI’ – in other words, when should users trust outputs from Artificial Intelligence (AI) models? GNS Science has been studying XAI as part of a collaboration with Callaghan Innovation since 2020. In the 2020/21 financial year, GNS Science concentrated on XAI approaches that can be applied to unsupervised AI and machine learning models. As part of the work in the 2020/21 financial year, graph models were identified as an area of growth in the realm of XAI, and these were chosen as a focus of the XAI effort in the 2021/22 financial year. (The author)
