Multi-resolution Annotations for Emoji Prediction

Abstract

Emojis are able to express various linguistic components, including emotions, sentiments, events, etc. Predicting the proper emojis associated with text provides a way to summarize the text accurately, and it has been proven to be a good auxiliary task to many Natural Language Understanding (NLU) tasks. Labels in existing emoji prediction datasets are all passage-based and are usually under the multi-class classification setting. However, in many cases, one single emoji cannot fully cover the theme of a piece of text. It is thus useful to infer the part of text related to each emoji. The lack of multi-label and aspectlevel emoji prediction datasets is one of the bottlenecks for this task. This paper annotates an emoji prediction dataset with passage-level multi-class/multi-label, and aspect-level multiclass annotations. We also present a novel annotation method with which we generate the aspect-level annotations. The annotations are generated heuristically, taking advantage of the self-attention mechanism in Transformer networks. We validate the annotations both automatically and manually to ensure their quality. We also benchmark the dataset with a pretrained BERT model.

An empirical study of emoji use in software development communication

“…Di↵erent studies have reported that by analyzing emoji usage, we could predict one's personality and tone, as emojis can enrich the message with additional information [46,73,29,82,40]. Several studies have analyzed emoji use, mainly focusing on how and in what kind of devices emojis are used [46,73,27], how emojis could be matched to the context in di↵erent levels and interpretations [75,31,68], and how demographic features might impact the emoji use [40]. Another thread of research investigated emoji use in instant messaging [29,99,51], and social media platforms like Twitter [82,96,87,54,38] and Facebook [93].…”

Making Your Tweets More Fancy: Emoji Insertion to Texts

“…Also, emoji prediction can be used in dialogue systems to recommend a suitable emoji (Xie et al, 2016), and using emojis in chatbot systems is effective to attract users, specifically in the conversation for mental wellbeing (Fadhil et al, 2018). Recently, Ma et al (2020) released an emoji label prediction dataset with passage-level multi-class/multi-label, and aspect-level multi-class annotations.…”