Profiling Irony Speech Spreaders on Social Networks Using Deep Cleaning and BERT.

چکیده مقاله

With irony, language is employed figuratively and subtly to mean the opposite of what is stated. In the case of sarcasm, a more aggressive type of irony, the intent is to mock or scorn a victim without excluding the possibility to hurt. Stereotypes are often used, especially in discussions about controversial issues such as immigration, sexism, and misogyny. Regarding PAN’s open submission toward tackling this issue, we use BERT (bidirectional encoder representations from transformers) as a way to identify ironic and sarcastic phrases from genuine ones in Twitter posts. Since the goal is to detect irony in texts published on social media, and usually social media users have a different writing style and their texts contain a variety of nonstandard language expressions, the input texts are deep cleaned before feeding into the BERT network. The experimental results show a significant improvement in the accuracy and training loss ratio of the BERT network by applying deep cleaning to input texts. Thus, we achieved up to 98.5 percent accuracy with the proposed method.