CKingCoder at SemEval-2023 Task 9: Multilingual Tweet Intimacy Analysis

24696

Authors

nan

Abstract

AbstractThe SemEval 2023 Task 9 Multilingual Tweet Intimacy Analysis, is a shared task for analysing the intimacy in the tweets posted on Twitter. The dataset was provided by Pei and Jurgens, who are part of the task organisers, for this task consists of tweets in various languages, such as Chinese, English, French, Italian, Portuguese, and Spanish. The testing dataset also had unseen languages such as Hindi, Arabic, Dutch and Korean. The tweets may or may not be related to intimacy. The task of our team was to score the intimacy in tweets and place it in the range of 05 based on the level of intimacy in the tweet using the dataset provided which consisted of tweets along with its scores. The intimacy score is used to indicate whether a tweet is intimate or not. Our team participated in the task and proposed the ROBERTa model to analyse the intimacy of the tweets.

Keywords

nan

Full Content

Proceedings of the The 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 2009–2013 July 13-14, 2023 ©2023 Association for Computational Linguistics CKingCoder at SemEval-2023 Task 9: Multilingual Tweet Intimacy Analysis Harish Kumar B, Naveen D, Prem B and Aarthi S (harishkumarb0711@gmail.com, dnaveen0502@gmail.com prembala1710@gmail.com, aarthigopinath.msec@gmail.com) Computer Science and Engineering, Meenakshi Sundararajan Engineering College Abstract The SemEval 2023 Task 9 Multilingual Tweet Intimacy Analysis , (Pei et al., 2023) is a shared task for analysing the intimacy in the tweets posted on Twitter. The dataset was provided by Pei and Jurgens (2020), who are part of the task organizers, for this task consists of tweets in various languages, such as Chinese, English, French, Italian, Portuguese, and Span- ish. The testing dataset also had unseen lan- guages such as Hindi, Arabic, Dutch and Ko- rean. The tweets may or may not be related to intimacy. The task provided was to score the in- timacy in tweets and place it in the range of 0–5 based on the level of intimacy in the tweet using the dataset provided which consisted of tweets along with its scores. The intimacy score is used to indicate whether a tweet is intimate or not. Our team participated in the task and pro- posed the ROBERTa model , Liu et al. (2019) to analyse the intimacy of the tweets. 1 Introduction Intimacy is the feeling that makes people feel close to each other. Intimacy is formed as a result of sharing knowledge and experience with each other. Intimacy is developed through a process of four phases, as per the Levinger and Snoek Interdepen- dence Model. The Four Phases of Intimacy are the No Contact Phase, the Awareness Phase, the Surface Contact Phase, and the Coexistence Phase. The no-contact phase is when the people involved don’t know each other. The awareness phase is when the people involved just know about each other but don’t have any superficial contact. The surface contact phase is when the people involved know each other and have superficial contact with each other. The coexistence phase is where the people involved have deep links and are mutually dependent on each other. Social media is now on an enormous upward curve. Intimacy is commonly referred to as physical intimacy. It has four major forms, among which one is physical intimacy. The other forms of intimacy are emotional intimacy, where intimacy forms personal bonds such as love or marriage; spiritual intimacy, where intimacy is formed due to the sharing of spiritual ideas or opin- ions; and intellectual intimacy, where intimacy is formed by shared common opinions between in- duividuals. In modern times, there has been a notable shift in the way individuals express their emotions in real-time. With the proliferation of social media, communication has largely moved online, with in- dividuals utilizing platforms such as Twitter, chat applications, and virtual meetings. Twitter, in par- ticular, has experienced tremendous growth, with over 500 million tweets being sent per day as of March 1, 2023. Users are able to share …

Content truncated to 500 words

Labeled Text

AbstractThe SemEval 2023 Task 9 Multilingual Tweet Intimacy Analysis, is a shared task for analysing the intimacy in the tweets posted on Twitter. The dataset was provided by Pei and Jurgens, who are part of the task organisers, for this task consists of tweets in various languages, such as Chinese, English, French, Italian, Portuguese, and Spanish. The testing dataset also had unseen languages such as Hindi, Arabic, Dutch and Korean. The tweets may or may not be related to intimacy. The task of our team was to score the intimacy in tweets and place it in the range of 05 based on the level of intimacy in the tweet using the dataset provided which consisted of tweets along with its scores. The intimacy score is used to indicate whether a tweet is intimate or not. Our team participated in the task and proposed the ROBERTa model to analyse the intimacy of the tweets.

Publication Details

Paper ID:
24696
Year:
2023
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan