Fangyu Liu (a student at Clare Hall, pictured) and Nigel Collier, together with collaborators from McGill University and the University of Copenhagen, have been awarded the Best Long Paper Award at the Empirical Methods in Natural Language Processing (EMNLP) Conference 2021.

The award-winning paper, ‘Visually grounded reasoning across languages and cultures’, aims to address the western-centric biases in existing resources for training and evaluating AI systems and provides a data construction protocol that is driven entirely by native speakers of a diverse range of languages. This sets up a new path forward for building AI systems that can effectively work for speakers of different languages and cultures around the world. 

Fangyu comments:
‘I am extremely honoured to receive the best paper award from EMNLP 2021. As a junior researcher, I am encouraged to keep working and contributing more to the community. Our awarded paper shows how current AI technologies fail to transfer across cultures and languages. My collaborators and I are determined to continue the work of making AI technologies more inclusive.’

For more information about this work you can read the paper at https://aclanthology.org/2021.emnlp-main.818.pdf