UIT-E10dot3 at SemEval-2021 Task 5: Toxic Spans Detection with Named Entity Recognition and Question-Answering Approaches

Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021), pages 919–926 Bangkok, Thailand (online), August 5–6, 2021. ©2021 Association for Computational Linguistics 919 UIT-E10dot3 at SemEval-2021 Task 5: Toxic Spans Detection with Named Entity Recognition and Question-Answering Approaches Phu Gia Hoang1, 2, Luan Thanh Nguyen1, 2, and Kiet Nguyen1, 2 1University of Information Technology, Ho Chi Minh City, Vietnam 2Vietnam National University Ho Chi Minh City, Vietnam {19520215, 17520721}@gm.uit.edu.vn, kietnv@uit.edu.vn Abstract The increment of toxic comments on online space is causing tremendous effects on other vulnerable users. For this reason, consider- able efforts are made to deal with this, and SemEval-2021 Task 5: Toxic Spans Detection is one of those. This task asks competitors to extract spans that have toxicity from the given texts, and we have done several analyses to understand its structure before doing experi- ments. We solve this task by two approaches, Named Entity Recognition with spaCy’s li- brary and Question-Answering with RoBERTa combining with ToxicBERT, and the former gains the highest F1-score of 66.99%. 1 Introduction The world of social media is overgrowing, and users easily express their opinions or feelings to- ward topics that they are concerned about. How- ever, because of the freedom of speech, lots of toxic comments or contents are uncontrollably increas- ing. There are several kinds of research about the effect of toxic speech on users’ health. In 2017, re- search about the impact of toxic language on health was conducted (Mohan et al., 2017). Sometimes, with toxic words, conversations can become cy- berbullying, cyber threats, or online harassment, which are harmful to users. To reduce those neg- ative impacts, there are abundant researches for classifying contents into toxic or non-toxic, and then they hide the whole text if it is toxic. However, that action may inhibit the freedom of speech. As a result, censoring only toxic spans is the better solu- tion for this problem. Therefore, in SemEval-2021 Task 5: Toxic Spans Detection (Pavlopoulos et al., 2021) we try to realize it. About toxic contents on the internet, researches were only about binary toxicity classification. Still, in task 5 of SemEval-2021, which is about toxic spans detection, we conduct more in-depth re- search into the toxicity, find exactly which parts of the text are toxic. As the NER approach and Question-Answering (QA) approach, we propose two approaches for solving this problem. We use RoBERTa (Liu et al., 2019) combining with Tox- icBERT (Hanu and Unitary team, 2020), transfer learning models, for QA approach and spaCy’s library (Honnibal and Montani, 2017) for NER ap- proach. We organize the paper as follows. Section 2 is related works that we consult for building the systems. The dataset and analyses are defined in Section 3. In section 4, we introduce our two pro- posed systems for toxic spans detection. Section 5 describes the results of the studies and analyses. Finally, in Section 6, we bring our work to a close. 2 Related Works Researchers around the world these days have started to concentrate …

Content truncated to 500 words