We are the Language and Multimodal Processing Group at the Department of Computer Science at the University of Copenhagen. Our recent work includes representation learning for multilingual and multimodal data, resource creation, and language modelling.
We are developing language models that can process any written language.
Retrieval augmentation of multimodal models for image captioning.
Design, understand, and evaluate multimodality in multilingual models
2024 |
FoodieQA: A Multimodal Dataset for Fine-Grained Understanding of Chinese Food Culture.
Proceedings of EMNLP. |
2024 |
Understanding Retrieval Robustness for Retrieval-augmented Image Captioning.
Proceedings of ACL. |
2024 |
Classification of Medical Text in Small and Imbalanced Datasets in a Non-English Language.
Proceedings of MIDL. |
2024 |
Compositional Generalization in Multimodal Models.
Proceedings of NAACL. |
2024 |
PAELLA: Parameter-Efficient Lightweight Language-Agnostic Captioning Model.
Findings of NAACL. |
2024 |
The Role of Data Curation in Image Captioning.
Proceedings of EACL. |