| Título |
DSVS at HOMO-MEX24: Multi-Class and Multi-Label Hate Speech Detection using Transformer-Based Models |
| Tipo |
Congreso |
| Sub-tipo |
Memoria |
| Descripción |
6th Iberian Languages Evaluation Forum, IberLEF 2024 |
| Resumen |
The present work describes the participation of the DSVS team in the HOMO-MEX shared task at IberLEF 2024 on detecting hate speech in online messages and music lyrics targeting the LGBTQ+ community, written in Mexican Spanish. The study addressed all three proposed tracks: Track 1 involves identifying LGBTQ+ categories (multiclass); Track 2 focuses on fine-grained hate speech detection (multi-labeled); and Track 3 involves homophobic lyrics detection (binary task). Through an exploration of the datasets, we employ various BERT-based models. Our team’s best submission secured the 4th position for Track 1, the 3rd position for Track 2, and the 9th position for Track 3. © 2024 Copyright for this paper by its authors. |
| Observaciones |
CEUR Workshop Proceedings, v. 3756 |
| Lugar |
Valladolid |
| País |
España |
| No. de páginas |
|
| Vol. / Cap. |
|
| Inicio |
2024-10-24 |
| Fin |
|
| ISBN/ISSN |
|
