Multilingual Eye-Tracking Corpus for Human and Machine-Based Language Processing

The aim of this project is the creation of a multi-lingual eye-tracking-while-reading corpus. In collaboration with our partners, we record eye movements on parallel texts from a range of typologically similar and dissimilar languages spoken in Europe (Basque, Danish, English, Estonian, French, Croatian, Latvian, Lithuanian, Maltese, Macedonian, Dutch, Portuguese, Russian, Slovenian). This multilingual reading corpus will serve as a resource for future research in psycholinguistics and computational linguistics. On the one hand, it can be used to cross-linguistically evaluate psycholinguistic theories on human language processing; on the other hand it can be used compare the linguistic behavior of machine-learning-based language models with human behavior, and, finally, may help to make these NLP systems exhibit a more human-like behavior.

This project is funded by the Foundation for Research in Science and the Humanities at the University of Zurich.

Project partners:

Clara Martin (Basque Center On Cognition, Brain and Language, Spain)
Itziar Gonzalez (University of the Basque Country, Spain)
Nora Hollenstein Dänemark (University of Copenhagen, Denmark)
Jens Roeser (Nottingham Trent University, UK)
Kaidi Lõo (University of Tartu, Estonia)
Lonneke van der Plas (Idiap, Martigny, Switzerland)
Marijan Palmovic, Kristina Cergol (University of Zagreb, Croatia)
Gunta Krumina (University of Latvia, Latvia)
Ramuné Kasperé (Kaunas University of Technology, Lithuania)
Holger Mitterer (University of Malta, Malta)
Veronika Kareva (South East European University, North Macedonia)
Stefan Frank (Radboud University, Nijmegen, The Netherlands)
João Veríssimo (University of Lisbon, Portugal)
Olga Dragoy (HSE University Moscow, Russia)
Artur Stepanov (University of Nova Gorica, Slovenia)
Anja Podlesek (University of Ljubljana, Slovenia)

Project duration: 2022-2025