Explainable Machine Learning for Scientific Insights and Discoveries

Jochen Garcke,Ribana Roscher,Marco F. Duarte,Bastian Bohn

doi:10.1109/access.2020.2976199

Jochen Garcke, Ribana Roscher

Open Access

https://doi.org/10.1109/access.2020.2976199

Copy DOI

Abstract

Machine learning methods have been remarkably successful for a wide range of application areas in the extraction of essential information from data. An exciting and relatively recent development is the uptake of machine learning in the natural sciences, where the major goal is to obtain novel scientific insights and discoveries from observational or simulated data. A prerequisite for obtaining a scientific outcome is domain knowledge, which is needed to gain explainability, but also to enhance scientific consistency. In this article we review explainable machine learning in view of applications in the natural sciences and discuss three core elements which we identified as relevant in this context: transparency, interpretability, and explainability. With respect to these core elements, we provide a survey of recent scientific works that incorporate machine learning and the way that explainable machine learning is used in combination with domain knowledge from the application areas.

Highlights

Machine learning methods, especially with the rise of neural networks (NNs), are nowadays used widely in commercial applications
In this work, we reviewed the concept of explainable machine learning and discerned between transparency, interpretability, and explainability
We discussed the possibility of influencing model design choices and the step of interpreting algorithmic outputs by domain knowledge and a posteriori consistency checks

Summary

INTRODUCTION

Especially with the rise of neural networks (NNs), are nowadays used widely in commercial applications. The three elements transparency, interpretability, and explainability play a central role These concepts will be defined and discussed in detail in this survey. Another essential component is domain knowledge, which is necessary to achieve explainability, but can be used to foster scientific consistency of the model and the result. While these terms are more methodology-driven and refer to properties of the model and the algorithm, we describe the role of additional information and domain knowledge, as well as scientific.

TERMINOLOGY

DISCUSSION