# Offensive Language Detection with BERT-based models, By Customizing Attention Probabilities

arXiv: https://arxiv.org/abs/2110.05133 · HTML: https://mlsanity.com/paper/2110.05133/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 41 references that are in the map.
- First posted on arXiv 11 Oct 2021.

## Builds on (most cited first, 21 of 21)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [Convolutional Neural Networks for Sentence Classification](https://mlsanity.com/paper/1408.5882/) (2014)
- [Universal Sentence Encoder](https://mlsanity.com/paper/1803.11175/) (2018)
- [Predicting the Type and Target of Offensive Posts in Social Media](https://mlsanity.com/paper/1902.09666/) (2019)
- [SemEval-2019 Task 6: Identifying and Categorizing Offensive Language in Social Media (OffensEval)](https://mlsanity.com/paper/1903.08983/) (2019)
- [Deep Learning for Hate Speech Detection in Tweets](https://mlsanity.com/paper/1706.00188/) (2017)
- [SemEval-2020 Task 12: Multilingual Offensive Language Identification in Social Media (OffensEval 2020)](https://mlsanity.com/paper/2006.07235/) (2020)
- [ParsBERT: Transformer-based Model for Persian Language Understanding](https://mlsanity.com/paper/2005.12515/) (2020)
- [KUISAIL at SemEval-2020 Task 12: BERT-CNN for Offensive Speech Identification in Social Media](https://mlsanity.com/paper/2007.13184/) (2020)
- [One-step and Two-step Classification for Abusive Language Detection on Twitter](https://mlsanity.com/paper/1706.01206/) (2017)
- [Detecting Hate Speech in Social Media](https://mlsanity.com/paper/1712.06427/) (2017)
- [Aggression-annotated Corpus of Hindi-English Code-mixed Data](https://mlsanity.com/paper/1803.09402/) (2018)
- [Developing a Multilingual Annotated Corpus of Misogyny and Aggression](https://mlsanity.com/paper/2003.07428/) (2020)
- [Medical Text Classification using Convolutional Neural Networks](https://mlsanity.com/paper/1704.06841/) (2017)
- [Kungfupanda at SemEval-2020 Task 12: BERT-Based Multi-Task Learning for Offensive Language Detection](https://mlsanity.com/paper/2004.13432/) (2020)
- [Galileo at SemEval-2020 Task 12: Multi-lingual Learning for Offensive Language Identification using Pre-trained Language Models](https://mlsanity.com/paper/2010.03542/) (2020)
- [BRUMS at SemEval-2020 Task 12 : Transformer based Multilingual Offensive Language Identification in Social Media](https://mlsanity.com/paper/2010.06278/) (2020)
- [UM-IU@LING at SemEval-2019 Task 6: Identifying Offensive Tweets Using BERT and SVMs](https://mlsanity.com/paper/1904.03450/) (2019)
- [GUIR at SemEval-2020 Task 12: Domain-Tuned Contextualized Models for Offensive Language Detection](https://mlsanity.com/paper/2007.14477/) (2020)

## Built on by (most cited first, 2 of 2)

- [Cyberbullying Detection for Low-resource Languages and Dialects: Review of the State of the Art](https://mlsanity.com/paper/2308.15745/) (2023)
- [OPSD: an Offensive Persian Social media Dataset and its baseline evaluations](https://mlsanity.com/paper/2404.05540/) (2024)

## Abstract (arXiv)

This paper describes a novel study on using `Attention Mask' input in transformers and using this approach for detecting offensive content in both English and Persian languages. The paper's principal focus is to suggest a methodology to enhance the performance of the BERT-based models on the `Offensive Language Detection' task. Therefore, we customize attention probabilities by changing the `Attention Mask' input to create more efficacious word embeddings. To do this, we firstly tokenize the training set of the exploited datasets (by BERT tokenizer). Then, we apply Multinomial Naive Bayes to map these tokens to two probabilities. These probabilities indicate the likelihood of making a text non-offensive or offensive, provided that it contains that token. Afterwards, we use these probabilities to define a new term, namely Offensive Score. Next, we create two separate (because of the differences in the types of the employed datasets) equations based on Offensive Scores for each language to re-distribute the `Attention Mask' input for paying more attention to more offensive phrases. Eventually, we put the F1-macro score as our evaluation metric and fine-tune several combinations of BERT with ANNs, CNNs and RNNs to examine the effect of using this methodology on various combinations. The results indicate that all models will enhance with this methodology. The most improvement was 2% and 10% for English and Persian languages, respectively.