Last week, ESMAS-ES+ team members Iván Arias, María José Domínguez and Carlos Valcárcel presented part of their research at the 21st EURALEX Congress, which took place in Cavtat (Croatia) from October 8th to 12th, 2024. Their poster, titled Sustainability and Lexicography: Evaluating the methodological approach for the development of an automatic semantic tagger, showcased recent advances in the analysis of the project’s underlying ontology to enhance semantic classes.

For this purpose, large language models (LLMs) like ChatGPT and Gemini were tested as annotators. ChatGPT achieved a precision rate of 70% when given specific instructions via few-shot prompting. However, polysemous lexical units still require further disambiguation.

The team was thrilled with the feedback from the expert audience at the congress, and we look forward to sharing more updates on the project’s progress!