DeepHateExplainer: Explainable Hate Speech Detection in Under-resourced Bengali Language
This work provides a method for explainable hate speech detection in Bengali, an under-resourced language, which is important for social media content moderation and user safety in Bengali-speaking communities. It is an incremental step in applying existing methods to new linguistic contexts.
This paper addresses hate speech detection in the under-resourced Bengali language by proposing DeepHateExplainer, a neural ensemble method combining transformer-based architectures. It achieves F1-scores of 78% for political, 91% for personal, 89% for geopolitical, and 84% for religious hate speech, outperforming traditional machine learning and deep neural network baselines.
The exponential growths of social media and micro-blogging sites not only provide platforms for empowering freedom of expressions and individual voices, but also enables people to express anti-social behaviour like online harassment, cyberbullying, and hate speech. Numerous works have been proposed to utilize textual data for social and anti-social behaviour analysis, by predicting the contexts mostly for highly-resourced languages like English. However, some languages are under-resourced, e.g., South Asian languages like Bengali, that lack computational resources for accurate natural language processing (NLP). In this paper, we propose an explainable approach for hate speech detection from the under-resourced Bengali language, which we called DeepHateExplainer. Bengali texts are first comprehensively preprocessed, before classifying them into political, personal, geopolitical, and religious hates using a neural ensemble method of transformer-based neural architectures (i.e., monolingual Bangla BERT-base, multilingual BERT-cased/uncased, and XLM-RoBERTa). Important(most and least) terms are then identified using sensitivity analysis and layer-wise relevance propagation(LRP), before providing human-interpretable explanations. Finally, we compute comprehensiveness and sufficiency scores to measure the quality of explanations w.r.t faithfulness. Evaluations against machine learning~(linear and tree-based models) and neural networks (i.e., CNN, Bi-LSTM, and Conv-LSTM with word embeddings) baselines yield F1-scores of 78%, 91%, 89%, and 84%, for political, personal, geopolitical, and religious hates, respectively, outperforming both ML and DNN baselines.