akaBERT at SemEval-2022 Task 6: An Ensemble Transformer-based Model for Arabic Sarcasm Detection

Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022), pages 885 - 890 July 14-15, 2022 ©2022 Association for Computational Linguistics akaBERT at SemEval-2022 Task 6: An Ensemble Transformer-based Model for Arabic Sarcasm Detection Abdulrahman Mohamed Kamr and Ensaf Hussein Mohamed Research Support Center in Computing and Informatics Department of Computer Science, Faculty of Computers and Artificial Intelligence, Helwan University, Cairo, Egypt. abdokamr94@fci.helwan.edu.eg, ensaf_hussein@fci.helwan.edu.eg Abstract Due to the widespread usage of social media sites and the enormous number of users who utilize irony implicit words in most of their tweets and posts, it has become necessary to detect sarcasm, which strongly influences understanding and analyzing the crowd’s opinions. Detecting sarcasm is difficult due to the nature of sarcastic tweets, which vary based on the topic, region, the user’s attitude, culture, terminologies, and other criteria. In addition to these difficulties, detecting sarcasm in Arabic has its challenges due to its complexities, such as being morphologically rich, having many different dialects, and having low resources. In this research, we present our submission of (iSarcasmEval) sub-task A of the shared task on SemEval 2022. In Sub-task A; we determine whether the tweets are sarcastic or non-sarcastic. We implemented different approaches based on Transformers. First, we fine-tuned the AraBERT, MARABERT, and AraELECTRA. One of the challenges that faced us was that the data was not balanced. Non-sarcastic data is much more than sarcastic. We used data augmentation techniques to balance the two classes, significantly affecting the performance. The performance F1 score of the three models was 87%, 90%, and 91%, respectively. Then we boosted the three models by developing an ensemble model based on hard voting. The final performance F1 Score was 93%. 1 Introduction According to the Cambridge dictionary definition of Sarcasm, it is the use of remarks that mean the opposite of what they say, made to hurt some- one’s feelings or humorously criticize something, Like when you say Love this weather. (When the weather is horrible). Sarcasm detection is determining whether or not a piece of text is sarcastic. Sarcasm is a significant challenge for sentiment analysis systems. This is because a sarcastic sentence usually contains an implicit negative sentiment that is expressed with positive expressions. This discrepancy between the surface and intended sentiments creates a difficult challenge for sentiment analysis systems. Sarcasm detection is a difficult task for a variety of reasons. First of all, there aren’t many labeled data resources for sarcasm detection. Moreover, any available texts that can be collected (for exam- ple, Tweets) contain many issues, such as an evolv- ing dictionary of slang words and abbreviations, so it usually takes many hours for human annotators to prepare the data for any potential use. Further- more, the nature of sarcasm detection adds to the task’s difficulty, as Sarcasm can be considered rela- tive and varies significantly between people, and it depends on many factors such as the topic, region, time, the events surrounding the sentence, and the readers/writers mentality and the; in other words, a sentence …

Content truncated to 500 words