Multilingual Twitter Corpus and Baselines for Evaluating Demographic Bias in Hate Speech Recognition

TL;DR

This paper introduces a multilingual Twitter dataset with inferred demographic attributes to evaluate bias in hate speech detection models across five languages, providing a new resource for fairness research.

Contribution

It presents a novel multilingual Twitter corpus with inferred demographic labels and benchmarks for assessing bias in hate speech classifiers.

Findings

Demographic labels are validated via crowdsourcing.

Baseline classifiers show varying bias levels across demographics.

Multilingual data reveals language-specific bias patterns.

Abstract

Existing research on fairness evaluation of document classification models mainly uses synthetic monolingual data without ground truth for author demographic attributes. In this work, we assemble and publish a multilingual Twitter corpus for the task of hate speech detection with inferred four author demographic factors: age, country, gender and race/ethnicity. The corpus covers five languages: English, Italian, Polish, Portuguese and Spanish. We evaluate the inferred demographic labels with a crowdsourcing platform, Figure Eight. To examine factors that can cause biases, we take an empirical analysis of demographic predictability on the English corpus. We measure the performance of four popular document classifiers and evaluate the fairness and bias of the baseline classifiers on the author-level demographic attributes.