uOttawa at SemEval-2018 Task 1: Self-Attentive Hybrid GRU-Based Network

Proceedings of the 12th International Workshop on Semantic Evaluation (SemEval-2018), pages 181–185 New Orleans, Louisiana, June 5–6, 2018. ©2018 Association for Computational Linguistics uOttawa at SemEval-2018 Task 1: Self-Attentive Hybrid GRU-Based Network Ahmed Husseini Orabi1, Mahmoud Husseini Orabi1, Diana Inkpen1, David Van Bruwaene2 1EECS, University of Ottawa, 800 King Edward Avenue, Ottawa, Canada 2VISR inc., 10 Dundas St E. Suite 600, Toronto, Canada {ahuss045, mhuss092, diana.inkpen}@uottawa.ca, d@visr.co Abstract We propose a novel attentive hybrid GRU- based network (SAHGN), which we used at SemEval-2018 Task 1: Affect in Tweets. Our network has two main characteristics, 1) has the ability to internally optimize its feature representation using attention mechanisms, and 2) provides a hybrid rep- resentation using a character-level Convo- lutional Neural Network (CNN), as well as a self-attentive word-level encoder. The key advantage of our model is its ability to sig- nify the relevant and important information that enables self-optimization. Results are reported on the valence intensity regression task. 1 Introduction Affect analysis is one of the main topics of nat- ural language processing (NLP). It involves many sub-tasks such as sentiment and valence analyses expressed in text. We focus on the task of determin- ing valence intensity. are Hand-crafted features and/or sentiment lexicons commonly used for affect analysis (Mohammad, Kiritchenko, & Zhu, 2013; Taboada, Brooke, Tofiloski, Voll, & Stede, 2011) with clas- sifiers such as random forest and support vector machines (SVM). Affect in tweets (AIT) is a challenging task as it requires handling an informal writing style, which typically has many grammar mistakes, slangs, and misspellings. In this paper, we present a self-attentive hybrid GRU-based network (SAHGN) that competed at SemEval-2018 Task 1 (Mohammad, Bravo- Marquez, Salameh, & Kiritchenko, 2018; Mohammad & Kiritchenko, 2018). Our contributions can be summarized as below. • The implementation of a social media text processor: A library to help process social media text such as short-forms, emoticons, emojis, misspellings, hash tags, and slangs, as well as tokenization, word normalization, and sentence encoding. • The implementation of a self-attentive deep learning system: This system can pre- dict valence and intensity with limited cor- pora and vocabulary, and yet can have ac- ceptable performance. 2 High-Level Description of Our System Our goal is to provide a system that can predict va- lence and intensity for short text. Figure 1 shows a high-level description of our solution, which con- sists of two main components, social media text processor (Section 3) and self-attentive hybrid GRU-based network (Section 4.2). Figure 1: System architecture. 3 Social Media Text Processor The social media text processor aims to provide a reliable and fast tokenization. It involves the fol- • lowing preprocessing steps: Use a named entity recognizer (NER) (Finkel, Grenager, & Manning, 2005) to identify enti- ties such as persons, names, and places, and then replace them accordingly. • Build a vocabulary using an NGram tokenizer. 181 • Tokenize sentences into a set of tokens, and then use them to encode text into a sequence of indices (Table 1), which are fed …

Content truncated to 500 words