Explainable AI

Abstract

With the availability of wide variety of complex machine learning tools, industries have started opting for modelling techniques which have high interpretability, leading to the development of a whole new branch in artific ial intelligence called Explainable AI. This project involves creating a quantitative measure to gauge the explaining ability of a model. We have performed literature survey to gain deeper understanding about the latest model development in the field. We have created a framework to benchmark the interpretability of various models, thus helping the industry in selecting the best model for their particular use case.