Explainable Neural Networks based on Additive Index Models

TL;DR

This paper introduces the Explainable Neural Network (xNN), a structured neural network designed to produce interpretable features and relationships, enhancing transparency in machine learning models.

Contribution

The paper presents a novel neural network architecture that facilitates interpretability by learning and extracting understandable features with regularization.

Findings

xNN provides interpretable features easily extracted from the network

Regularization yields parsimonious explanations of feature-output relationships

Demonstrated effectiveness on simulated examples

Abstract

Machine Learning algorithms are increasingly being used in recent years due to their flexibility in model fitting and increased predictive performance. However, the complexity of the models makes them hard for the data analyst to interpret the results and explain them without additional tools. This has led to much research in developing various approaches to understand the model behavior. In this paper, we present the Explainable Neural Network (xNN), a structured neural network designed especially to learn interpretable features. Unlike fully connected neural networks, the features engineered by the xNN can be extracted from the network in a relatively straightforward manner and the results displayed. With appropriate regularization, the xNN provides a parsimonious explanation of the relationship between the features and the output. We illustrate this interpretable feature--engineering property on simulated examples.