Hrafn Loftsson
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hrafn Loftsson has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Andy Way
Dublin City University, Dublin, Ireland
2 shared publications - Steinthor Steingrimsson
Department of Computer Science Reykjavik University Iceland
2 shared publications - Anna Björk Nikulásdóttir1 shared publication
- Anton Karl Ingason
University of Iceland
1 shared publication Benedikt Geir Jóhannesson
Reykjavik University,Dept. of Computer Science,Reykjavik,Iceland
1 shared publication- Einar Freyr Sigurðsson
The Árni Magnusson Institute for Icelandic Studies
1 shared publication
Latest Publications
- Open Access
Aligning Language Models for Icelandic Legal Text Summarization
The integration of language models in the legal domain holds considerable promise for streamlining processes and improving efficiency in managing extensive workloads. However, the specialized terminology, nuanced language, and formal style of legal texts can present substantial challenges. This study examines whether preference-based training techniques, specifically Reinforcement Learning from Human Feedback and Direct Preference Optimization, can enhance models' performance in generating Icelandic legal summaries that align with domain-specific language standards and user preferences. We compare models fine-tuned with preference training to those using conventional supervised learning. Results indicate that preference training improves the legal accuracy of generated summaries over standard fine-tuning but does not significantly enhance the overall quality of Icelandic language usage. Discrepancies between automated metrics and human evaluations further underscore the importance of qualitative assessment in developing language models for the legal domain.
Open AccessAligning Language Models for Icelandic Legal Text Summarization
The integration of language models in the legal domain holds considerable promise for streamlining processes and improving efficiency in managing extensive workloads. However, the specialized terminology, nuanced language, and formal style of legal texts can present substantial challenges. This study examines whether preference-based training techniques, specifically Reinforcement Learning from Human Feedback and Direct Preference Optimization, can enhance models' performance in generating Icelandic legal summaries that align with domain-specific language standards and user preferences. We compare models fine-tuned with preference training to those using conventional supervised learning. Results indicate that preference training improves the legal accuracy of generated summaries over standard fine-tuning but does not significantly enhance the overall quality of Icelandic language usage. Discrepancies between automated metrics and human evaluations further underscore the importance of qualitative assessment in developing language models for the legal domain.
- Open Access
SentAlign: Accurate and Scalable Sentence Alignment
We present SentAlign, an accurate sentence alignment tool designed to handle very large parallel document pairs. Given user-defined parameters, the alignment algorithm evaluates all possible alignment paths in fairly large documents of thousands of sentences and uses a divide-and-conquer approach to align documents containing tens of thousands of sentences. The scoring function is based on LaBSE bilingual sentence representations. SentAlign outperforms five other sentence alignment tools when evaluated on two different evaluation sets, German-French and English-Icelandic, and on a downstream machine translation task.
Open AccessSentAlign: Accurate and Scalable Sentence Alignment
We present SentAlign, an accurate sentence alignment tool designed to handle very large parallel document pairs. Given user-defined parameters, the alignment algorithm evaluates all possible alignment paths in fairly large documents of thousands of sentences and uses a divide-and-conquer approach to align documents containing tens of thousands of sentences. The scoring function is based on LaBSE bilingual sentence representations. SentAlign outperforms five other sentence alignment tools when evaluated on two different evaluation sets, German-French and English-Icelandic, and on a downstream machine translation task.
- 26 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 318-324
Effective Entity Disambiguation in Low-Resource Languages: A Study of Icelandic
Entity disambiguation (ED) is integral to the task of entity linking (EL), the task of identifying named entities in text and linking them to their corresponding entries in a knowledge base (KB). In this paper, we present an effective and efficient ED system for Icelandic, using the Icelandic Wikipedia as a KB. We focus on disambiguation, the linking aspect of EL, assuming that an entity mention has already been located. We perform candidate generation using an alias table and Wikipedia search, and achieve candidate ranking through fine-grained entity typing and the use of an entity-aware Icelandic language model, IceLUKE. We study and compare the effects of different variations of candidate generation and candidate ranking, with the best approach reaching an accuracy of 95.2%. Our results highlight the importance of using an entity-aware language model in the candidate ranking step and show a minor improvement in using fine-grained entity typing to decrease the size of the candidate set before ranking.
- 26 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 318-324
Effective Entity Disambiguation in Low-Resource Languages: A Study of Icelandic
Entity disambiguation (ED) is integral to the task of entity linking (EL), the task of identifying named entities in text and linking them to their corresponding entries in a knowledge base (KB). In this paper, we present an effective and efficient ED system for Icelandic, using the Icelandic Wikipedia as a KB. We focus on disambiguation, the linking aspect of EL, assuming that an entity mention has already been located. We perform candidate generation using an alias table and Wikipedia search, and achieve candidate ranking through fine-grained entity typing and the use of an entity-aware Icelandic language model, IceLUKE. We study and compare the effects of different variations of candidate generation and candidate ranking, with the best approach reaching an accuracy of 95.2%. Our results highlight the importance of using an entity-aware language model in the candidate ranking step and show a minor improvement in using fine-grained entity typing to decrease the size of the candidate set before ranking.
- Open Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 256-263
SentAlign: Accurate and Scalable Sentence Alignment
Steinthor Steingrimsson, Hrafn Loftsson, Andy Way. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: System Demonstrations. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 256-263
Open AccessSentAlign: Accurate and Scalable Sentence Alignment
Steinthor Steingrimsson, Hrafn Loftsson, Andy Way. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: System Demonstrations. 2023.
- Open Access
- 1 January 2021
- Published by Association for Computational Linguistics (ACL)
- p. 9-14
IceSum: An Icelandic Text Summarization Corpus
Jón Daðason, Hrafn Loftsson, Salome Sigurðardóttir, Þorsteinn Björnsson. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Student Research Workshop. 2021.
- 1 January 2021
- Published by Association for Computational Linguistics (ACL)
- p. 9-14
Open AccessIceSum: An Icelandic Text Summarization Corpus
Jón Daðason, Hrafn Loftsson, Salome Sigurðardóttir, Þorsteinn Björnsson. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Student Research Workshop. 2021.
- 26 September 2020
- Published by Springer Nature
- p. 46-57
Named Entity Recognition for Icelandic: Annotated Corpus and Models
- 26 September 2020
- Published by Springer Nature
- p. 46-57
Named Entity Recognition for Icelandic: Annotated Corpus and Models
- 1 September 2020
- Published by Springer Nature
- p. 95-103
Experimenting with Different Machine Translation Models in Medium-Resource Settings
- 1 September 2020
- Published by Springer Nature
- p. 95-103
Experimenting with Different Machine Translation Models in Medium-Resource Settings
- Open Access
Language Technology Programme for Icelandic 2019-2023
In this paper, we describe a new national language technology programme for Icelandic. The programme, which spans a period of five years, aims at making Icelandic usable in communication and interactions in the digital world, by developing accessible, open-source language resources and software. The research and development work within the programme is carried out by a consortium of universities, institutions, and private companies, with a strong emphasis on cooperation between academia and industries. Five core projects will be the main content of the programme: language resources, speech recognition, speech synthesis, machine translation, and spell and grammar checking. We also describe other national language technology programmes and give an overview over the history of language technology in Iceland.
Open AccessLanguage Technology Programme for Icelandic 2019-2023
In this paper, we describe a new national language technology programme for Icelandic. The programme, which spans a period of five years, aims at making Icelandic usable in communication and interactions in the digital world, by developing accessible, open-source language resources and software. The research and development work within the programme is carried out by a consortium of universities, institutions, and private companies, with a strong emphasis on cooperation between academia and industries. Five core projects will be the main content of the programme: language resources, speech recognition, speech synthesis, machine translation, and spell and grammar checking. We also describe other national language technology programmes and give an overview over the history of language technology in Iceland.
- 1 June 2008
- Published by Cambridge University Press (CUP) in Nordic Journal of Linguistics
- Vol. 31 (1) , 47-72
Tagging Icelandic text: A linguistic rule-based approach
The Icelandic language is a morphologically complex language, for which a large tagset has been created. This paper describes the design of a linguistic rule-based system for part-of-speech tagging Icelandic text. The system contains two main components: a disambiguator, IceTagger, and an unknown word guesser, IceMorphy. IceTagger uses a small number of local elimination rules along with a global heuristics component. The heuristics guess the functional roles of the words in a sentence, mark prepositional phrases, and use the acquired knowledge to force feature agreement where appropriate. IceMorphy is used for guessing the tag profile for unknown words and for automatically filling tag profile gaps in the lexicon. Evaluation shows that IceTagger achieves 91.54% accuracy, a substantial improvement on the highest accuracy, 90.44%, obtained using three state-of-the-art data-driven taggers. Furthermore, the accuracy increases to 92.95% by using IceTagger along with two data-driven taggers in a simple voting scheme. The development time of the tagging system was only seven man-months, which can be considered a short development period for a linguistic rule-based system.
- 1 June 2008
- Published by Cambridge University Press (CUP) in Nordic Journal of Linguistics
- Vol. 31 (1) , 47-72
Tagging Icelandic text: A linguistic rule-based approach
The Icelandic language is a morphologically complex language, for which a large tagset has been created. This paper describes the design of a linguistic rule-based system for part-of-speech tagging Icelandic text. The system contains two main components: a disambiguator, IceTagger, and an unknown word guesser, IceMorphy. IceTagger uses a small number of local elimination rules along with a global heuristics component. The heuristics guess the functional roles of the words in a sentence, mark prepositional phrases, and use the acquired knowledge to force feature agreement where appropriate. IceMorphy is used for guessing the tag profile for unknown words and for automatically filling tag profile gaps in the lexicon. Evaluation shows that IceTagger achieves 91.54% accuracy, a substantial improvement on the highest accuracy, 90.44%, obtained using three state-of-the-art data-driven taggers. Furthermore, the accuracy increases to 92.95% by using IceTagger along with two data-driven taggers in a simple voting scheme. The development time of the tagging system was only seven man-months, which can be considered a short development period for a linguistic rule-based system.