Nothing Else Matters: Model-Agnostic Explanations By Identifying Prediction Invariance

TL;DR

This paper introduces anchor-LIME, a model-agnostic explanation method that provides high-precision, rule-based explanations with clear coverage boundaries, improving interpretability across various domains.

Contribution

The paper presents anchor-LIME, a novel model-agnostic explanation technique that offers high-precision, rule-based explanations with well-defined coverage boundaries, enhancing interpretability.

Findings

aLIME outperforms linear LIME in simulated experiments

aLIME provides flexible, high-precision explanations across domains

Clear coverage boundaries improve interpretability

Abstract

At the core of interpretable machine learning is the question of whether humans are able to make accurate predictions about a model's behavior. Assumed in this question are three properties of the interpretable output: coverage, precision, and effort. Coverage refers to how often humans think they can predict the model's behavior, precision to how accurate humans are in those predictions, and effort is either the up-front effort required in interpreting the model, or the effort required to make predictions about a model's behavior. In this work, we propose anchor-LIME (aLIME), a model-agnostic technique that produces high-precision rule-based explanations for which the coverage boundaries are very clear. We compare aLIME to linear LIME with simulated experiments, and demonstrate the flexibility of aLIME with qualitative examples from a variety of domains and tasks.