Presentation Schedule
Automated Detection of Hate Speech and Toxic Comments Using Machine Learning and Natural Language Processing (96204)
Session: On Demand
Room: Virtual Video Presentation
Presentation Type:Virtual Presentation
Hate speech and toxic comments have become a big difficulty in online communication, leading to societal issues such as cyberbullying, discrimination, and misinformation. This study offers an automated approach for detecting hate speech and poisonous comments using Natural Language Processing (NLP) and Machine Learning (ML). The proposed approach makes use of transformer-based models like BERT and RoBERTa, which have been tested on the Jigsaw Toxic Comment dataset. The study compares classical ML models (logistic regression, naïve Bayes) against deep learning-based approaches to assess their effectiveness. The results reveal that transformer-based models beat typical ML models in terms of precision, recall, and F1 score, indicating their usefulness in detecting context-aware hate speech. The research also explores the ethical concerns of hate speech identification, as well as the issues of bias in dataset categorization.
Authors:
Aman Deep Singh, Nirma University, India
Dhruvesh Vaghasiya, Nirma University, India
Dev Detroja, Nirma University, India
Vedant Vaghasiya, Nirma University, India
About the Presenter(s)
Dr Aman Deep Singh is a University Assistant Professor/Lecturer at Nirma University in India
See this presentation on the full schedule – On Demand Schedule
Comments
Powered by WP LinkPress