Natural Language Processing Lab
Major Competencies
Specialized in the development and training of expansive language models using high-performance distributed computing environments. Rooted at the University of Iceland, there's a distinctive emphasis on collaborative research, partnering with entities like Miðeind and other startups to foster innovation and knowledge dissemination in the field of NLP.
Focus on Icelandic NLP and Development.
Our lab at the University of Iceland is centered on natural language processing research and the development of language technology for Icelandic, in collaboration with various companies and institutions.
Icelandic NLP as a paradigm.
We build upon the foundations laid by the Icelandic language technology program, while also dedicating efforts towards the creation of new language resources and solutions.
Training High-Performance Models.
A significant area of our expertise lies in training models within high-performance distributed computing environments, which is crucial for advancing our research and developing effective language technology solutions.
Lab news and events
People
Prof. Dr. Hafsteinn Einarsson
Hafsteinn is an assistant professor at the School of Engineering and Natural Sciences of the University of Iceland. He received his Ph.D. in Computer Science from ETH in 2017. He has worked on applied ML solutions for startups and in the Icelandic banking sector. He is currently focused on natural language processing, interpretable ML methods and optimization problems.
Annika Simonsen
Annika is an Ph.D. student in artificial intelligence and language technology at the School of Engineering and Natural Sciences of the University of Iceland. She had an MSc in Applied Linguistics from the University of Edinburgh ('19) and an MA in Language Technology from the University of Iceland ('24). Her Ph.D. is part of the TrustLLM project, which is developing an open, trustworthy, and sustainable LLM initially targeting the Germanic languages. Annika's Ph.D. project revolves around the alignment of the neural network that will be developed in the TrustLLM project with a special focus on the low-resource languages.
Hans Erik Mathias Stenlund
Mathias is a Ph.D. student in the fields of language technology and high-performance computing at the School of Engineering and Natural Sciences of the University of Iceland. He is currently partaking in the European TrustLLM project that aims to create the next generation of trustworthy and open LLMs for the Germanic languages. In 2023, he earned his master’s degree in language technology from Uppsala University, where he previously also earned his bachelor’s degree in linguistics.
Vésteinn Snæbjarnarson
Vésteinn, a dedicated ELLIS PhD student at the Pioneer Centre for Artificial Intelligence, delves into the realms of natural language processing and computer vision. Based at the University of Copenhagen and advised by Serge Belongie, with co-advisory from Ryan Cotterell at ETH Zürich, he explores multimodal settings that combine methods for NLP and Computer Vision. His research pursuits also encompass compositionality of embedding spaces and generative models. Vésteinn's academic journey commenced with a BA in Philosophy and a BS in Mathematics from the University of Iceland, followed by an MS in Computer Science, where his thesis addressed Question Answering for Icelandic. He's also associated with Icelandic language technology company, Miðeind ehf. Currently, Vésteinn's work pivots around descriptive image captioning and fine-grained visual categorization.
Projects & Cooperations
All IHPC Projects
Selected Publications
Journals
Friðriksdóttir, Steinunn Rut; Einarsson, Hafsteinn (2022): Friðriksdóttir & Einarsson (2022), Bootstrapping Icelandic Knowledge Graph Data. University of Amsterdam / Amsterdam University of Applied Sciences. Conference contribution. https://doi.org/10.21942/uva.20367948.v1
Friðriksdóttir, S. R., Eggertsson, V. Á., Jóhannesson, B. G., Daníelsson, H., Loftsson, H., & Einarsson, H. (2022, June). Building an Icelandic Entity Linking Corpus. In Proceedings of the Workshop on Dataset Creation for Lower-Resourced Languages within the 13th Language Resources and Evaluation Conference (pp. 27-35).
Friðriksdóttir, S. R., & Einarsson, H. (2022, June). Fictionary-Based Games for Language Resource Creation. In Proceedings of the 2nd Workshop on Novel Incentives in Data Collection from People: models, implementations, challenges and results within LREC 2022 (pp. 25-31).
Snæbjarnarson, V., & Einarsson, H. (2022, July). Cross-Lingual QA as a Stepping Stone for Monolingual Open QA in Icelandic. In Proceedings of the Workshop on Multilingual Information Access (MIA) (pp. 29-36).
Snæbjarnarson, V., & Einarsson, H. (2022, June). Natural questions in icelandic. In Proceedings of the Thirteenth Language Resources and Evaluation Conference (pp. 4488-4496).
Snæbjarnarson, V., Símonarson, H. B., Ragnarsson, P. O., Ingólfsdóttir, S. L., Jónsson, H., Þorsteinsson, V., & Einarsson, H. (2022, June). A Warm Start and a Clean Crawled Corpus-A Recipe for Good Language Models. In Proceedings of the Thirteenth Language Resources and Evaluation Conference (pp. 4356-4366).