THU_NGN at SemEval-2018 Task 3: Tweet Irony Detection with Densely connected LSTM and Multi-task Learning

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 51–56 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics THU NGN at SemEval-2018 Task 3: Tweet Irony Detection with Densely Connected LSTM and Multi-task Learning Chuhan Wu1, Fangzhao Wu2, Sixing Wu1, Junxin Liu1, Zhigang Yuan1 and Yongfeng Huang1 1Tsinghua National Laboratory for Information Science and Technology, Department of Electronic Engineering, Tsinghua University Beijing 100084, China 2Microsoft Research Asia {wuch15,wu-sx15,ljx16,yuanzg14,yfhuang}@mails.tsinghua.edu.cn wufangzhao@gmail.com Abstract Detecting irony is an important task to mine fine-grained information from social web mes- sages. Therefore, the Semeval-2018 task 3 is aimed to detect the ironic tweets (subtask A) and their irony types (subtask B). In or- der to address this task, we propose a system based on a densely connected LSTM network with multi-task learning strategy. In our dense LSTM model, each layer will take all outputs from previous layers as input. The last LSTM layer will output the hidden representations of texts, and they will be used in three classifi- cation task. In addition, we incorporate sev- eral types of features to improve the model performance. Our model achieved an F-score of 70.54 (ranked 2/43) in the subtask A and 49.47 (ranked 3/29) in the subtask B. The ex- perimental results validate the effectiveness of our system. 1 Introduction Figurative languages such as irony are widely used in web messages such as tweets to convey different sentiment. Identifying the ironic texts can help to understand the social web better and has many ap- plications such as sentiment analysis (Ghosh and Veale, 2016). Irony detecting techniques are im- portant to improve the performance of sentiment analysis. For example, the tweet “Monday morn- ings are my fave:)# not” is an irony with nega- tive sentiment, but it will be probably classified as a positive one by a standard sentiment analysis model (Van Hee et al., 2016b). Thus, capturing the ironic information in texts is useful to predict sentiment more accurately (Van Hee et al., 2016a). However, determining whether a text is ironic is a challenging task since the the differences be- tween ironic and non-ironic texts are usually sub- tle. For example, the tweet “Love this weather #not” is ironic, but a similar tweet “Hate this weather #not happy” is non-ironic. Different approaches are proposed to recognize the com- plex irony in texts. Existing methods to detect irony are mainly based on rules or machine learn- ing techniques (Joshi et al., 2017). Rules based methods usually depend on lexicons to identify irony (Khattri et al., 2015; Maynard and Green- wood, 2014). However, these methods cannot uti- lize the contextual information from texts. Tra- ditional machine learning based methods such as SVM (Desai and Dave, 2016) are also effective in this task, but they usually need manually fea- ture engineering (Barbieri et al., 2014). Recently, deep learning techniques are successfully applied to this task. For example, Ghosh et al. (2016) propose to use a CNN-LSTM model to classify the ironic and non-ironic tweets. Their method can …

Content truncated to 500 words