Concept-Based Explanations to Test for False Causal Relationships Learned by Abusive Language Classifiers

TL;DR

This paper investigates whether abusive language classifiers learn false causal links with concepts like negative emotions, proposing explanation metrics to detect over-reliance and false causality in models.

Contribution

It introduces concept-based explanation metrics to identify and compare false causal relationships learned by classifiers, especially when challenge sets are unavailable.

Findings

Classifiers often learn false causal links with negative emotions.

Explanation metrics can detect over-reliance on specific concepts.

Methods enable comparison of models regarding concept influence.

Abstract

Classifiers tend to learn a false causal relationship between an over-represented concept and a label, which can result in over-reliance on the concept and compromised classification accuracy. It is imperative to have methods in place that can compare different models and identify over-reliances on specific concepts. We consider three well-known abusive language classifiers trained on large English datasets and focus on the concept of negative emotions, which is an important signal but should not be learned as a sufficient feature for the label of abuse. Motivated by the definition of global sufficiency, we first examine the unwanted dependencies learned by the classifiers by assessing their accuracy on a challenge set across all decision thresholds. Further, recognizing that a challenge set might not always be available, we introduce concept-based explanation metrics to assess the influence of the concept on the labels. These explanations allow us to compare classifiers regarding the degree of false global sufficiency they have learned between a concept and a label.