midr: Learning from Black-Box Models by Maximum Interpretation Decomposition

TL;DR

The paper introduces the R package midr, which implements Maximum Interpretation Decomposition to create interpretable additive surrogate models from black-box models, aiding explainability in machine learning.

Contribution

It presents a novel functional decomposition method, MID, and its implementation in an R package for interpreting black-box models through additive approximations.

Findings

Provides a new tool for global surrogate modeling.

Enables detailed interpretation of complex black-box models.

Demonstrates practical application and features of the midr package.

Abstract

The use of appropriate methods of Interpretable Machine Learning (IML) and eXplainable Artificial Intelligence (XAI) is essential for adopting black-box predictive models in fields where model and prediction explainability is required. As a novel tool for interpreting black-box models, we introduce the R package midr, which implements Maximum Interpretation Decomposition (MID). MID is a functional decomposition approach that derives a low-order additive representation of a black-box model by minimizing the squared error between the model's prediction function and this additive representation. midr enables learning from black-box models by constructing a global surrogate model with advanced analytical capabilities. After reviewing related work and the theoretical foundation of MID, we demonstrate the package's usage and discuss some of its key features.