Cover Image

A Parallel Cross-Lingual Benchmark for Multimodal Idiomaticity Understanding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Torunoğlu-Selamet, Dilara, Arslan, Dogukan, Wilkens, Rodrigo, He, Wei, Eryiğit, Doruk, Pickard, Thomas, Pagano, Adriana S., Villavicencio, Aline, Eryiğit, Gülşen, Abuczki, Ágnes, Cardoso, Aida, Lazarenka, Alesia, Almassova, Dina, Mendes, Amalia, Kanellopoulou, Anna, Brosa-Rodríguez, Antoni, Saulite, Baiba, Wojtowicz, Beata, Pedersen, Bolette, Hidalgo-Ternero, Carlos Manuel, Liebeskind, Chaya, Jokić, Danka, Alves, Diego, Triantafyllidi, Eleni, Velldal, Erik, Philippy, Fred, Oleskeviciene, Giedre Valunaite, Rizgeliene, Ieva, Skadina, Inguna, Lobzhanidze, Irina, Haugen, Isabell Stinessen, Krito, Jauza Akbar, Marković, Jelena M., Monti, Johanna, Sauca, Josue Alejandro, Dobrovoljc, Kaja, Ugwuanyi, Kingsley O., Rituma, Laura, Øvrelid, Lilja, Agro, Maha Tufail, Abjalova, Manzura, Chatzigrigoriou, Maria, Ramos, María del Mar Sánchez, Pendevska, Marija, Seyyedrezaei, Masoumeh, Shamsfard, Mehrnoush, Ahsan, Momina, Khan, Muhammad Ahsan Riaz, Norman, Nathalie Carmen Hau, Ayyıldız, Nilay Erdem, Hosseini-Kivanani, Nina, Ligeti-Nagy, Noémi, Naeem, Numaan, Kanishcheva, Olha, Yatsyshyna, Olha, Orel, Daniil, Giommarelli, Petra, Osenova, Petya, Garabik, Radovan, Semou, Regina E., Rebechi, Rozane, Pranida, Salsabila Zahirah, Touileb, Samia, Nimb, Sanni, Ahmad, Sarfraz, Sharipova, Sarvinoz, Golan, Shahar, Ji, Shaoxiong, Aboh, Sopuruchi Christian, Sucur, Srdjan, Markantonatou, Stella, Olsen, Sussi, Tajalli, Vahide, Lipp, Veronika, Giouli, Voula, Eraydın, Yelda Yeşildal, Saaberi, Zahra, Xie, Zhuohan
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866910030959214592
author	Torunoğlu-Selamet, Dilara Arslan, Dogukan Wilkens, Rodrigo He, Wei Eryiğit, Doruk Pickard, Thomas Pagano, Adriana S. Villavicencio, Aline Eryiğit, Gülşen Abuczki, Ágnes Cardoso, Aida Lazarenka, Alesia Almassova, Dina Mendes, Amalia Kanellopoulou, Anna Brosa-Rodríguez, Antoni Saulite, Baiba Wojtowicz, Beata Pedersen, Bolette Hidalgo-Ternero, Carlos Manuel Liebeskind, Chaya Jokić, Danka Alves, Diego Triantafyllidi, Eleni Velldal, Erik Philippy, Fred Oleskeviciene, Giedre Valunaite Rizgeliene, Ieva Skadina, Inguna Lobzhanidze, Irina Haugen, Isabell Stinessen Krito, Jauza Akbar Marković, Jelena M. Monti, Johanna Sauca, Josue Alejandro Dobrovoljc, Kaja Ugwuanyi, Kingsley O. Rituma, Laura Øvrelid, Lilja Agro, Maha Tufail Abjalova, Manzura Chatzigrigoriou, Maria Ramos, María del Mar Sánchez Pendevska, Marija Seyyedrezaei, Masoumeh Shamsfard, Mehrnoush Ahsan, Momina Khan, Muhammad Ahsan Riaz Norman, Nathalie Carmen Hau Ayyıldız, Nilay Erdem Hosseini-Kivanani, Nina Ligeti-Nagy, Noémi Naeem, Numaan Kanishcheva, Olha Yatsyshyna, Olha Orel, Daniil Giommarelli, Petra Osenova, Petya Garabik, Radovan Semou, Regina E. Rebechi, Rozane Pranida, Salsabila Zahirah Touileb, Samia Nimb, Sanni Ahmad, Sarfraz Sharipova, Sarvinoz Golan, Shahar Ji, Shaoxiong Aboh, Sopuruchi Christian Sucur, Srdjan Markantonatou, Stella Olsen, Sussi Tajalli, Vahide Lipp, Veronika Giouli, Voula Eraydın, Yelda Yeşildal Saaberi, Zahra Xie, Zhuohan
author_facet	Torunoğlu-Selamet, Dilara Arslan, Dogukan Wilkens, Rodrigo He, Wei Eryiğit, Doruk Pickard, Thomas Pagano, Adriana S. Villavicencio, Aline Eryiğit, Gülşen Abuczki, Ágnes Cardoso, Aida Lazarenka, Alesia Almassova, Dina Mendes, Amalia Kanellopoulou, Anna Brosa-Rodríguez, Antoni Saulite, Baiba Wojtowicz, Beata Pedersen, Bolette Hidalgo-Ternero, Carlos Manuel Liebeskind, Chaya Jokić, Danka Alves, Diego Triantafyllidi, Eleni Velldal, Erik Philippy, Fred Oleskeviciene, Giedre Valunaite Rizgeliene, Ieva Skadina, Inguna Lobzhanidze, Irina Haugen, Isabell Stinessen Krito, Jauza Akbar Marković, Jelena M. Monti, Johanna Sauca, Josue Alejandro Dobrovoljc, Kaja Ugwuanyi, Kingsley O. Rituma, Laura Øvrelid, Lilja Agro, Maha Tufail Abjalova, Manzura Chatzigrigoriou, Maria Ramos, María del Mar Sánchez Pendevska, Marija Seyyedrezaei, Masoumeh Shamsfard, Mehrnoush Ahsan, Momina Khan, Muhammad Ahsan Riaz Norman, Nathalie Carmen Hau Ayyıldız, Nilay Erdem Hosseini-Kivanani, Nina Ligeti-Nagy, Noémi Naeem, Numaan Kanishcheva, Olha Yatsyshyna, Olha Orel, Daniil Giommarelli, Petra Osenova, Petya Garabik, Radovan Semou, Regina E. Rebechi, Rozane Pranida, Salsabila Zahirah Touileb, Samia Nimb, Sanni Ahmad, Sarfraz Sharipova, Sarvinoz Golan, Shahar Ji, Shaoxiong Aboh, Sopuruchi Christian Sucur, Srdjan Markantonatou, Stella Olsen, Sussi Tajalli, Vahide Lipp, Veronika Giouli, Voula Eraydın, Yelda Yeşildal Saaberi, Zahra Xie, Zhuohan
contents	Potentially idiomatic expressions (PIEs) construe meanings inherently tied to the everyday experience of a given language community. As such, they constitute an interesting challenge for assessing the linguistic (and to some extent cultural) capabilities of NLP systems. In this paper, we present XMPIE, a parallel multilingual and multimodal dataset of potentially idiomatic expressions. The dataset, containing 34 languages and over ten thousand items, allows comparative analyses of idiomatic patterns among language-specific realisations and preferences in order to gather insights about shared cultural aspects. This parallel dataset allows to evaluate model performance for a given PIE in different languages and whether idiomatic understanding in one language can be transferred to another. Moreover, the dataset supports the study of PIEs across textual and visual modalities, to measure to what extent PIE understanding in one modality transfers or implies in understanding in another modality (text vs. image). The data was created by language experts, with both textual and visual components crafted under multilingual guidelines, and each PIE is accompanied by five images representing a spectrum from idiomatic to literal meanings, including semantically related and random distractors. The result is a high-quality benchmark for evaluating multilingual and multimodal idiomatic language understanding.
format	Preprint
id	arxiv_https___arxiv_org_abs_2601_08645
institution	arXiv
publishDate	2026
record_format	arxiv
spellingShingle	A Parallel Cross-Lingual Benchmark for Multimodal Idiomaticity Understanding Torunoğlu-Selamet, Dilara Arslan, Dogukan Wilkens, Rodrigo He, Wei Eryiğit, Doruk Pickard, Thomas Pagano, Adriana S. Villavicencio, Aline Eryiğit, Gülşen Abuczki, Ágnes Cardoso, Aida Lazarenka, Alesia Almassova, Dina Mendes, Amalia Kanellopoulou, Anna Brosa-Rodríguez, Antoni Saulite, Baiba Wojtowicz, Beata Pedersen, Bolette Hidalgo-Ternero, Carlos Manuel Liebeskind, Chaya Jokić, Danka Alves, Diego Triantafyllidi, Eleni Velldal, Erik Philippy, Fred Oleskeviciene, Giedre Valunaite Rizgeliene, Ieva Skadina, Inguna Lobzhanidze, Irina Haugen, Isabell Stinessen Krito, Jauza Akbar Marković, Jelena M. Monti, Johanna Sauca, Josue Alejandro Dobrovoljc, Kaja Ugwuanyi, Kingsley O. Rituma, Laura Øvrelid, Lilja Agro, Maha Tufail Abjalova, Manzura Chatzigrigoriou, Maria Ramos, María del Mar Sánchez Pendevska, Marija Seyyedrezaei, Masoumeh Shamsfard, Mehrnoush Ahsan, Momina Khan, Muhammad Ahsan Riaz Norman, Nathalie Carmen Hau Ayyıldız, Nilay Erdem Hosseini-Kivanani, Nina Ligeti-Nagy, Noémi Naeem, Numaan Kanishcheva, Olha Yatsyshyna, Olha Orel, Daniil Giommarelli, Petra Osenova, Petya Garabik, Radovan Semou, Regina E. Rebechi, Rozane Pranida, Salsabila Zahirah Touileb, Samia Nimb, Sanni Ahmad, Sarfraz Sharipova, Sarvinoz Golan, Shahar Ji, Shaoxiong Aboh, Sopuruchi Christian Sucur, Srdjan Markantonatou, Stella Olsen, Sussi Tajalli, Vahide Lipp, Veronika Giouli, Voula Eraydın, Yelda Yeşildal Saaberi, Zahra Xie, Zhuohan Computation and Language Potentially idiomatic expressions (PIEs) construe meanings inherently tied to the everyday experience of a given language community. As such, they constitute an interesting challenge for assessing the linguistic (and to some extent cultural) capabilities of NLP systems. In this paper, we present XMPIE, a parallel multilingual and multimodal dataset of potentially idiomatic expressions. The dataset, containing 34 languages and over ten thousand items, allows comparative analyses of idiomatic patterns among language-specific realisations and preferences in order to gather insights about shared cultural aspects. This parallel dataset allows to evaluate model performance for a given PIE in different languages and whether idiomatic understanding in one language can be transferred to another. Moreover, the dataset supports the study of PIEs across textual and visual modalities, to measure to what extent PIE understanding in one modality transfers or implies in understanding in another modality (text vs. image). The data was created by language experts, with both textual and visual components crafted under multilingual guidelines, and each PIE is accompanied by five images representing a spectrum from idiomatic to literal meanings, including semantically related and random distractors. The result is a high-quality benchmark for evaluating multilingual and multimodal idiomatic language understanding.
title	A Parallel Cross-Lingual Benchmark for Multimodal Idiomaticity Understanding
topic	Computation and Language
url	https://arxiv.org/abs/2601.08645