Interpretable Regional Descriptors: Hyperbox-Based Local Explanations

TL;DR

This paper presents a new hyperbox-based method for generating interpretable local explanations of machine learning predictions, providing understandable 'if-then' style arguments and feature influence insights.

Contribution

It introduces IRDs, a novel hyperbox-based approach for local, model-agnostic explanations, formalizes their computation as an optimization problem, and unifies existing methods within a new framework.

Findings

Benchmark results identify the best IRD strategies.

Two methods improve the quality of IRDs.

IRDs effectively explain feature influences and biases.

Abstract

This work introduces interpretable regional descriptors, or IRDs, for local, model-agnostic interpretations. IRDs are hyperboxes that describe how an observation's feature values can be changed without affecting its prediction. They justify a prediction by providing a set of "even if" arguments (semi-factual explanations), and they indicate which features affect a prediction and whether pointwise biases or implausibilities exist. A concrete use case shows that this is valuable for both machine learning modelers and persons subject to a decision. We formalize the search for IRDs as an optimization problem and introduce a unifying framework for computing IRDs that covers desiderata, initialization techniques, and a post-processing method. We show how existing hyperbox methods can be adapted to fit into this unified framework. A benchmark study compares the methods based on several quality measures and identifies two strategies to improve IRDs.