Research Repository

Emojional: Emoji Embeddings

Barry, Elena and Jameel, Shoaib and Raza, Haider (2021) Emojional: Emoji Embeddings. In: UK Workshop on Computational Intelligence (UKCI) 2021, 2021-09-08 - 2021-09-10, Aberystwyth, UK. (In Press)

[img]
Preview
Text
Emojinal_Submitted_Copy.pdf - Accepted Version

Download (1MB) | Preview

Abstract

Emojis are intended to illustrate or replace words in natural language. Although they are understandable across linguistic barriers, their implications are ever-evolving, and change depending on who we are talking to, and when. The underlying emotion is key. This is evident in language online, highlighted recently by the rise of cancel culture via online shaming. For example, the use of the clown emoji to signify someone is making a fool of themself, or the collective spamming of the snake emoji to “cancel” someone, both show seemingly innocent emojis being used as clear forms of aggression online. To capture these nuances, we have created novel emoji embeddings trained on their emotional content. The subsequent emoji embeddings are generally more accurate than the state-of-the-art embeddings on the task of sentiment analysis. These embeddings can be found in our GitHub repository: https://github.com/elenabarry/emojional.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Published proceedings: _not provided_
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Depositing User: Elements
Date Deposited: 22 Jul 2021 15:07
Last Modified: 22 Sep 2021 13:15
URI: http://repository.essex.ac.uk/id/eprint/30781

Actions (login required)

View Item View Item