The objective of this task is to predict one or more labels for each tweet containing LGBT+phobic hate speech. The labels represent different types of phobias: lesbophobia, directed at homosexual women; gayphobia, directed at homosexual men; biphobia, against people attracted to more than one gender; transphobia, directed at non-cisgender people; and other LGBT+phobias, which includes hatred towards sexual and gender minorities not covered by the previous categories, such as aphobia. The label Unrelated to LGBT+ (NR) is also included for tweets not related to this community.
Publication
Gómez-Adorno et al. (2024). Overview of HOMO-MEX at IberLEF 2024: Hate Speech Detection Towards the Mexican Spanish speaking LGBT+ Population. Procesamiento del Lenguaje Natural, Revista, 73: 393-405.
Language
Spanish (Mexico)
NLP topic
Dataset
Year
2024
Publication link
Task results
| System | MacroF1 Sort ascending |
|---|---|
| CANTeam | 0.9730 |
| Homomex | 0.9488 |
| DSVS | 0.9436 |
| VerbaNexAI | 0.9393 |
| UC-CUJAE | 0.9346 |
