publications.bib

@inproceedings{yang-etal-2024-elco,
    title = "The {ELC}o Dataset: Bridging Emoji and Lexical Composition",
    author = "Yang, Zi Yun  and
      Zhang, Ziqing  and
      Miao, Yisong",
    editor = "Calzolari, Nicoletta  and
      Kan, Min-Yen  and
      Hoste, Veronique  and
      Lenci, Alessandro  and
      Sakti, Sakriani  and
      Xue, Nianwen",
    booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)",
    month = may,
    year = "2024",
    address = "Torino, Italia",
    publisher = "ELRA and ICCL",
    url = "https://aclanthology.org/2024.lrec-main.1381/",
    pages = "15899--15909",
    abstract = "Can emojis be composed to convey intricate meanings like English phrases? As a pioneering study, we present the Emoji-Lexical Composition (ELCo) dataset, a new resource that offers parallel annotations of emoji sequences corresponding to English phrases. Our dataset contains 1,655 instances, spanning 209 diverse concepts from tangible ones like {\textquotedblleft}right man{\textquotedblright} (✔️👨) to abstract ones such as {\textquotedblleft}full attention{\textquotedblright} (🧐✍️, illustrating a metaphoric composition of a focusing face and writing hand). ELCo enables the analysis of the patterns shared between emoji and lexical composition. Through a corpus study, we discovered that simple strategies like direct representation and reduplication are sufficient for conveying certain concepts, but a richer, metaphorical strategy is essential for expressing more abstract ideas. We further introduce an evaluative task, Emoji-based Textual Entailment (EmoTE), to assess the proficiency of NLP models in comprehending emoji compositions. Our findings reveals the challenge of understanding emoji composition in a zero-shot setting for current models, including ChatGPT. Our analysis indicates that the intricacy of metaphorical compositions contributes to this challenge. Encouragingly, models show marked improvement when fine-tuned on the ELCo dataset, with larger models excelling in deciphering nuanced metaphorical compositions."
}