Hate Speech Detection and Racial Bias Mitigation in Social Media based on BERT model

Abstract

Disparate biases associated with datasets and trained classifiers in hateful and abusive content identification tasks have raised many concerns recently. Although the problem of biased datasets on abusive language detection has been addressed more frequently, biases arising from trained classifiers have not yet been a matter of concern. Here, we first introduce a transfer learning approach for hate speech detection based on an existing pre-trained language model called BERT and evaluate the proposed model on two publicly available datasets annotated for racism, sexism, hate or offensive content on Twitter. Next, we introduce a bias alleviation mechanism in hate speech detection task to mitigate the effect of bias in training set during the fine-tuning of our pre-trained BERT-based model. Toward that end, we use an existing regularization method to reweight input samples, thereby decreasing the effects of high correlated training set' s n-grams with class labels, and then fine-tune our pre-trained BERT-based model with the new re-weighted samples. To evaluate our bias alleviation mechanism, we employ a cross-domain approach in which we use the trained classifiers on the aforementioned datasets to predict the labels of two new datasets from Twitter, AAE-aligned and White-aligned groups, which indicate tweets written in African-American English (AAE) and Standard American English (SAE) respectively. The results show the existence of systematic racial bias in trained classifiers as they tend to assign tweets written in AAE from AAE-aligned group to negative classes such as racism, sexism, hate, and offensive more often than tweets written in SAE from White-aligned. However, the racial bias in our classifiers reduces significantly after our bias alleviation mechanism is incorporated. This work could institute the first step towards debiasing hate speech and abusive language detection systems.

arXiv commentThis paper has been accepted in the PLOS ONE journal in August 2020

References in corpus (4)

• Efficient Estimation of Word Representations in Vector Space
• Stereotypical Bias Removal for Hate Speech Detection Task using Knowledge-based Generalizations
• Transfer Learning for Hate Speech Detection in Social Media
• Analyzing Right-wing YouTube Channels: Hate, Violence and Discrimination

Cited by in corpus (15)

• Interactive Model Cards: A Human-Centered Approach to Model Documentation
• SoK: Content Moderation in Social Media, from Guidelines to Enforcement, and Research to Practice
• Introducing MBIB -- the first Media Bias Identification Benchmark Task and Dataset Collection
• Towards generalisable hate speech detection: a review on obstacles and solutions
• Lost in Moderation: How Commercial Content Moderation APIs Over- and Under-Moderate Group-Targeted Hate Speech and Linguistic Variations
• Character-level HyperNetworks for Hate Speech Detection
• Challenges in Automated Debiasing for Toxic Language Detection
• Sexism Prediction in Spanish and English Tweets Using Monolingual and Multilingual BERT and Ensemble Models
• Hate Speech Classifiers Learn Human-Like Social Stereotypes
• Automatic Sexism Detection with Multilingual Transformer Models
• Specializing General-purpose LLM Embeddings for Implicit Hate Speech Detection across Datasets
• Values That Are Explicitly Present in Fairy Tales: Comparing Samples from German, Italian and Portuguese Traditions
• A Novel Information-Theoretic Objective to Disentangle Representations for Fair Classification
• Battling Hateful Content in Indic Languages HASOC '21
• MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets