Fairness for Whom? Understanding the Reader's Perception of Fairness in Text Summarization

TL;DR

This paper investigates how readers perceive fairness in text summaries, revealing that perceptions are context-sensitive and that standard metrics like ROUGE do not capture perceived fairness, proposing new human-in-the-loop and automated metrics.

Contribution

It introduces a human-in-the-loop and graph-based automated methodology to quantify perceived fairness in summaries, addressing a gap in existing fairness evaluation methods.

Findings

Reader's fairness perception varies with context.

ROUGE scores do not reflect perceived fairness.

Proposed metrics effectively quantify perceived bias.

Abstract

With the surge in user-generated textual information, there has been a recent increase in the use of summarization algorithms for providing an overview of the extensive content. Traditional metrics for evaluation of these algorithms (e.g. ROUGE scores) rely on matching algorithmic summaries to human-generated ones. However, it has been shown that when the textual contents are heterogeneous, e.g., when they come from different socially salient groups, most existing summarization algorithms represent the social groups very differently compared to their distribution in the original data. To mitigate such adverse impacts, some fairness-preserving summarization algorithms have also been proposed. All of these studies have considered normative notions of fairness from the perspective of writers of the contents, neglecting the readers' perceptions of the underlying fairness notions. To bridge this gap, in this work, we study the interplay between the fairness notions and how readers perceive them in textual summaries. Through our experiments, we show that reader's perception of fairness is often context-sensitive. Moreover, standard ROUGE evaluation metrics are unable to quantify the perceived (un)fairness of the summaries. To this end, we propose a human-in-the-loop metric and an automated graph-based methodology to quantify the perceived bias in textual summaries. We demonstrate their utility by quantifying the (un)fairness of several summaries of heterogeneous socio-political microblog datasets.