Hybrid Deep Learning Type-2 Fuzzy Logic Systems For Explainable AI
Ravikiran Chimatapu, Hani Hagras, Mathias Kern, Gilbert Owusu
Abstract
The recent years have witnessed a rapid rise in the use of Artificial Intelligence (AI) systems, in particular Machine Learning (ML) models. The vast majority of AI systems employ black box models that lack transparency in operation and decision making. This lack of transparency curtails the use of these AI systems in regulated applications (such as medical, financial applications, etc.) where it is important to understand the reasoning behind the predictions of the AI system. In these situations, interpretable models need to be used. However, interpretable models can turn into black-box models for high dimensional inputs. There are a variety of approaches that have been proposed to solve this problem. In this paper, we present a novel hybrid deep learning type-2 fuzzy logic system for explainable AI which addresses these challenges to provide a highly interpretable model that has reasonable performance when compared to the other black box models.