Results of the summarizing process for the arXiv paper: 2111.14830v1

The paper titled "Abusive and Threatening Language Detection in Urdu using Boosting based and BERT based models: A Comparative Approach" addresses the growing concern of online hatred on social media platforms. To tackle this issue, various social media platforms have implemented moderation policies and employed moderators to identify and take action against posts that violate these policies. Additionally, researchers in the field of abusive language detection have conducted studies to improve the detection of such content. While there has been extensive research on abusive language detection in English, there is a significant gap in detecting abusive language in low resource languages like Hindi and Urdu. The organizers of this task provided a dataset for detecting abusive language and threatening content specifically in Urdu. In their study, the authors explored several machine learning models including XGboost, LGBM, and m-BERT based models for abusive and threatening content detection in Urdu. They observed that a Transformer model trained on an abusive language dataset in Arabic yielded the best performance. The results of their study showed that their model achieved first place for both abusive and threatening content detection with F1 scores of 0.88 and 0.54 respectively. Overall, this paper highlights the importance of addressing online hatred by developing effective methods for detecting abusive language, particularly in low resource languages like Urdu. The authors' comparative approach utilizing different machine learning models demonstrates promising results for identifying abusive and threatening content on social media platforms.

• - The paper addresses the concern of online hatred on social media platforms
• - Various social media platforms have implemented moderation policies and employed moderators to tackle this issue
• - There is a significant gap in detecting abusive language in low resource languages like Hindi and Urdu
• - The authors explored machine learning models including XGboost, LGBM, and m-BERT for abusive and threatening content detection in Urdu
• - A Transformer model trained on an abusive language dataset in Arabic yielded the best performance
• - The model achieved first place for both abusive and threatening content detection with F1 scores of 0.88 and 0.54 respectively
• - Effective methods for detecting abusive language, particularly in low resource languages like Urdu, are important in addressing online hatred

The paper talks about how people say mean things on social media. Some social media platforms are trying to stop this by having rules and hiring people to check the bad comments. But it's harder to find mean words in languages like Hindi and Urdu. The authors tried different computer models to find mean and scary words in Urdu, and a model trained with Arabic words worked the best. This model got the highest scores for finding mean and scary words. It's important to have good ways of finding mean words, especially in languages like Urdu, so we can stop online hatred." Definitions- Online hatred: When people say mean things to each other on the internet. - Social media platforms: Websites or apps where people can share things with others, like Facebook or Twitter. - Moderation policies: Rules that social media platforms have to make sure people don't say bad things. - Machine learning models: Computer programs that can learn from examples and make decisions on their own. - Abusive language: Words or phrases that are meant to hurt or insult someone. - Threatening content: Messages or posts that make someone feel scared or unsafe. - Transformer model: A type of machine learning model that is good at understanding language.

Abusive and Threatening Language Detection in Urdu Using Boosting Based and BERT Based Models: A Comparative Approach

Social media platforms have become a major source of communication for people all over the world. Unfortunately, these platforms are also becoming increasingly used to spread hate speech, abusive language, and threatening content. To address this growing concern, various social media platforms have implemented moderation policies and employed moderators to identify posts that violate these policies. Additionally, researchers in the field of abusive language detection have conducted studies to improve the detection of such content. While there has been extensive research on abusive language detection in English, there is a significant gap in detecting abusive language in low resource languages like Hindi and Urdu. The paper titled "Abusive and Threatening Language Detection in Urdu using Boosting based and BERT based models: A Comparative Approach" addresses this issue by exploring several machine learning models for detecting abusive language specifically in Urdu.

Dataset

The organizers of this task provided a dataset for detecting abusive language and threatening content specifically in Urdu. This dataset was divided into two categories: Abusive (A) and Threatening (T). Each category contained 1000 tweets labeled as either “abusive” or “not-abusive”/”threatening” or “not-threatening” respectively.

Models Explored

In their study, the authors explored several machine learning models including XGboost, LGBM, m-BERT based models for both A & T categories separately with different hyperparameters settings to compare their performance on the given dataset. They observed that a Transformer model trained on an Arabic dataset yielded the best performance when tested against their own datasets for both A & T categories separately with F1 scores of 0.88 & 0.54 respectively which was better than other boosting based algorithms used by them e..g XGboost & LGBM etc .

Conclusion

Overall, this paper highlights the importance of addressing online hatred by developing effective methods for detecting abusive language particularly in low resource languages like Urdu through comparative approach utilizing different machine learning models which demonstrates promising results for identifying abusive and threatening content on social media platforms .