DeepAI AI Chat
Log In Sign Up

Listening to Affected Communities to Define Extreme Speech: Dataset and Experiments

by   Antonis Maronikolakis, et al.

Building on current work on multilingual hate speech (e.g., Ousidhoum et al. (2019)) and hate speech reduction (e.g., Sap et al. (2020)), we present XTREMESPEECH, a new hate speech dataset containing 20,297 social media passages from Brazil, Germany, India and Kenya. The key novelty is that we directly involve the affected communities in collecting and annotating the data - as opposed to giving companies and governments control over defining and combatting hate speech. This inclusive approach results in datasets more representative of actually occurring online speech and is likely to facilitate the removal of the social media content that marginalized communities view as causing the most harm. Based on XTREMESPEECH, we establish novel tasks with accompanying baselines, provide evidence that cross-country training is generally not feasible due to cultural differences between countries and perform an interpretability analysis of BERT's predictions.


page 7

page 15


HASOCOne@FIRE-HASOC2020: Using BERT and Multilingual BERT models for Hate Speech Detection

Hateful and Toxic content has become a significant concern in today's wo...

HateCheckHIn: Evaluating Hindi Hate Speech Detection Models

Due to the sheer volume of online hate, the AI and NLP communities have ...

Detecting Hate Speech in Social Media

In this paper we examine methods to detect hate speech in social media, ...

CRAB: Class Representation Attentive BERT for Hate Speech Identification in Social Media

In recent years, social media platforms have hosted an explosion of hate...

The Effect of Extremist Violence on Hateful Speech Online

User-generated content online is shaped by many factors, including endog...

From Senseless Swarms to Smart Mobs: Tuning Networks for Prosocial Behaviour

Social media have been seen to accelerate the spread of negative content...