UIT-E10dot3 at SemEval-2021 Task 5: Toxic Spans Detection with Named Entity Recognition and Question-Answering Approaches
This work addresses toxic comment detection for online platforms, presenting an incremental improvement with specific performance gains.
The paper tackled the problem of detecting toxic spans in online comments by using Named Entity Recognition and Question-Answering approaches, achieving a highest F1-score of 66.99% with the NER method.
The increment of toxic comments on online space is causing tremendous effects on other vulnerable users. For this reason, considerable 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 experiments. We solve this task by two approaches, Named Entity Recognition with spaCy library and Question-Answering with RoBERTa combining with ToxicBERT, and the former gains the highest F1-score of 66.99%.