Publication Details

Category Text Publication
Reference Category Book chapters
DOI 10.18653/v1/2024.climatenlp-1.7
Licence creative commons licence
Title (Primary) Using LLMs to build a database of climate extreme impacts
Title (Secondary) Proceedings of the 1st Workshop on Natural Language Processing Meets Climate Change (ClimateNLP 2024), Bangkok, Thailand, August 16, 2024
Author Li, N.; Zahra, S.; de Brito, M.M.; Flynn, C.M.; Görnerup, O.; Worou, K.; Kurfalı, M.; Meng, C.; Thiery, W.; Zscheischler, J. ORCID logo ; Messori, G.; Nivre, J.
Publisher Stammbach, D.; Ni, J.; Schimanski, T.; Dutia, K.; Singh, A.; Bingler, J.; Christiaen, C.; Kushwaha, N.; Muccione, V.; Vaghefi, S.A.; Leippold, M.
Year 2024
Department SUSOZ; CER
Page From 93
Page To 110
Language englisch
Topic T5 Future Landscapes
Abstract To better understand how extreme climate,events impact society, we need to increase the,availability of accurate and comprehensive in-,formation about these impacts. We propose a,method for building large-scale databases of,climate extreme impacts from online textual,sources, using LLMs for information extraction,in combination with more traditional NLP tech-,niques to improve accuracy and consistency.,We evaluate the method against a small bench-,mark database created by human experts and,find that extraction accuracy varies for different,types of information. We compare three differ-,ent LLMs and find that, while the commercial,GPT-4 model gives the best performance over-,all, the open-source models Mistral and Mixtral,are competitive for some types of information.
Persistent UFZ Identifier https://www.ufz.de/index.php?en=20939&ufzPublicationIdentifier=29522
Li, N., Zahra, S., de Brito, M.M., Flynn, C.M., Görnerup, O., Worou, K., Kurfalı, M., Meng, C., Thiery, W., Zscheischler, J., Messori, G., Nivre, J. (2024):
Using LLMs to build a database of climate extreme impacts
In: Stammbach, D., Ni, J., Schimanski, T., Dutia, K., Singh, A., Bingler, J., Christiaen, C., Kushwaha, N., Muccione, V., Vaghefi, S.A., Leippold, M. (eds.)
Proceedings of the 1st Workshop on Natural Language Processing Meets Climate Change (ClimateNLP 2024), Bangkok, Thailand, August 16, 2024
Association for Computational Linguistics, Kerrville, p. 93 - 110 10.18653/v1/2024.climatenlp-1.7