Papers by Hugo Sousa
Text2Story Lusa: A Dataset for Narrative Analysis in European Portuguese News Articles (2024.lrec-main)
Copied to clipboard
Sérgio Nunes, Alípio Mario Jorge, Evelin Amorim, Hugo Sousa, António Leal, Purificação Moura Silvano, Inês Cantante, Ricardo Campos
| Challenge: | Access to annotated corpora with narrative elements is limited due to the lack of readily available datasets and copyright concerns. |
| Approach: | They developed a dataset that contains 357 news articles and 117 manually annotated articles with over 50 thousand individual annotations. |
| Outcome: | The proposed datasets are available in English and Portuguese and are based on 117 articles totaling over 50 thousand individual annotations. |
Puntuguese: A Corpus of Puns in Portuguese with Micro-edits (2024.lrec-main)
Copied to clipboard
Marcio Lima Inacio, Gabriela Wick-Pedro, Renata Ramisch, Luís Espírito Santo, Xiomara S. Q. Chacon, Roney Santos, Rogério Sousa, Rafael Anchiêta, Hugo Goncalo Oliveira
| Challenge: | Existing corpus of punning humor in Portuguese is unfit for machine learning due to data leakage. |
| Approach: | They propose to use Puntuguese to create a corpus of punning humor in Portuguese that is significantly more difficult to recognize than the previous corpus. |
| Outcome: | The proposed corpus achieves an F1-Score of 68.9% and is significantly more difficult than the previous corpus. |