Tracks

TRACK 1: Multi-dialect polarity detection track (Multi-class)

In this track, participants will be given Reddit posts with LGBTQ+ keywords. They will have to indicate the polarity (positive, negative, neutral) of the post towards the keyword. 

In this subtask, the Spanish dialects of the training are the same as those in the evaluation. Specifically, Reddit posts from Argentina, Chile, Colombia, and Mexico will be used for both training and evaluation.

The keywords are: lexicon: trans, lgbt, gay/gays, lesbiana/as, bisexual/es, asexual/es, transexual/les, travesti/is, queer/s, transgénero, pansexual/es, intersexual/es

TRACK 2: Cross-dialect polarity detection (Multi-labeled)

In this track, participants will be given Reddit posts with LGBTQ+ keywords. They will have to indicate the post’s polarity (positive, negative, neutral) towards the keyword. In this subtask, the Spanish dialects of the training are not the same as those in evaluation.

In this track, the evaluation set will contain posts from Argentina, Chile, Colombia, and Mexico. The posts of the test dataset will belong to Bolivia, Costa Rica, Cuba, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, Panama, Paraguay, Peru, Puerto Rico, Uruguay, and Venezuela.