AdMIRe: Advancing Multimodal Idiomaticity Representation (SemEval-2025 Task 1) - Labelled Datasets

<p dir="ltr">The <b>AdMIRe </b>shared task was organised and run as SemEval-2025 Task 1: <a href="https://semeval.github.io/SemEval2025/" rel="noreferrer" target="_blank">https://semeval.github.io/SemEval2025/</a></p><p><br></p><p dir="ltr">The datasets contain potentially idiomatic expressions (PIEs) in English (EN) and Brazilian Portuguese (PT), context sentences in which the expressions are used in either a literal or idiomatic sense and associated images depicting the expressions with either a single image or a sequence of three images capturing change over time (like a comic strip).</p><p><br></p><p dir="ltr">See the task website (<a href="https://semeval2025-task1.github.io/" rel="noreferrer" target="_blank">https://semeval2025-task1.github.io/</a>), the attached task description document (SemEval_2025_Task_1__ADMIRE___Advancing_Multimodal_Idiomaticity_Representation.pdf) or the following task paper for more information:</p><p dir="ltr">Thomas Pickard, Aline Villavicencio, Maggie Mi, Wei He, Dylan Phelps, Carolina Scarton and Marco Idiart. 2025. <b>SemEval-2025 Task 1: AdMIRe - Advancing Multimodal Idiomaticity Representation.</b> <i>Proceedings of the 19th International Workshop on Semantic Evaluations (SemEval-2025). Association for Computational Linguistics, Vienna, Austria.</i></p>