Irony Detection in Persian Language: A Transfer Learning Approach Using Emoji Prediction

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2839–2845 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2839 Irony Detection in Persian Language: A Transfer Learning Approach Using Emoji Prediction Preni Golazizian, Behnam Sabeti, Seyed Arad Ashrafi Asli, Zahra Majdabadi, Omid Momenzadeh and Reza Fahmi Miras Technologies International Number 92, Movahed Danesh St., Tehran, Iran {preni, behnam, arad, zahra, omid, reza}@miras-tech.com Abstract Irony is a linguistic device used to intend an idea while articulating an opposing expression. Many text analytic algorithms used for emotion extraction or sentiment analysis, produce invalid results due to the use of irony. Persian speakers use this device more often due to the language’s nature and some cultural reasons. This phenomenon also appears in social media platforms such as Twitter where users express their opinions using ironic or sarcastic posts. In the current research, which is the first attempt at irony detection in Persian language, emoji prediction is used to build a pretrained model. The model is finetuned utilizing a set of hand-labeled tweets with irony tags. A bidirectional LSTM (BiLSTM) network is employed as the basis of our model which is improved by attention mechanism. Additionally, a Persian corpus for irony detection containing 4339 manually-labeled tweets is introduced. Experiments show the proposed approach outperforms the adapted state-of-the-art method tested on Persian dataset with an accuracy of 83.1%, and offers a strong baseline for further research in Persian language. Keywords: Irony Detection, Emoji Prediction, Transfer Learning, Multitask Learning, Deep Learning 1. Introduction Irony is a creative phenomenon that has been widely stud- ied in linguistics, philosophy, psychology and cognitive sci- ence (Colston and Gibbs, 2007), however, it’s difficult to reach a consensus on a single general definition for it. Most theorists would agree that in irony the literal meaning of the words does not hold and the speaker says something that seems to be the opposite of what they mean (Colston and Gibbs, 2007). In the current research sarcasm is considered as a special case of irony (Far´ıas et al., 2016; Gibbs, 2000). In recent years, Twitter has turned into a source of informa- tion about users’ expressions, ideas and opinions almost in any domain; therefore it has gained a lot of interest among researchers and companies who aim at user sentiment anal- ysis, opinion mining, emotion recognition and other similar tasks. Twitter includes a high percentage of tweets with irony us- age among users. Persian speakers use this figurative term even more often due to the Persian language’s nature and some cultural reasons. Presence of irony in a text can flip its sentiment’s polarity, thus irony-aware predictions are cru- cial for more accurate performance. The majority of the researches in automatic irony detec- tion task has been addressed in English. The lack of such models for the Persian language has motivated us for the current research. This study aimed at irony detection in Persian language which was the first attempt to the best of our …

Content truncated to 500 words