Counterfactual Visual Explanations

TL;DR

This paper introduces a method for generating counterfactual visual explanations by identifying and replacing regions in images to show how the system's predicted class can change, enhancing interpretability and human understanding.

Contribution

The work presents a novel technique for producing counterfactual visual explanations that improve interpretability and aid human learning in image classification tasks.

Findings

Counterfactual explanations improve human ability to distinguish bird species.

The method provides qualitative insights into model decision boundaries.

Users trained with explanations perform better in classification tasks.

Abstract

In this work, we develop a technique to produce counterfactual visual explanations. Given a 'query' image $I$ for which a vision system predicts class $c$, a counterfactual visual explanation identifies how $I$ could change such that the system would output a different specified class $c'$. To do this, we select a 'distractor' image $I'$ that the system predicts as class $c'$ and identify spatial regions in $I$ and $I'$ such that replacing the identified region in $I$ with the identified region in $I'$ would push the system towards classifying $I$ as $c'$. We apply our approach to multiple image classification datasets generating qualitative results showcasing the interpretability and discriminativeness of our counterfactual explanations. To explore the effectiveness of our explanations in teaching humans, we present machine teaching experiments for the task of fine-grained bird classification. We find that users trained to distinguish bird species fare better when given access to counterfactual explanations in addition to training examples.