HUrtful HUmour (HUHU): Detection of humour spreading prejudice in Twitter

In HUHU, our focus is on examining automatically the use of humor to express prejudice towards minorities, specifically analyzing Spanish tweets that are prejudicial towards:

Women and feminists

LGBTIQ community

Immigrants and racially discriminated people

Overweight people

For this, we propose 3 subtasks :

1-HUrtful HUmour Detection:

The first subtask consists in determining whether a prejudicial tweet is intended to cause humour. Participants will have to distinguish between tweets that using humour express prejudice and tweets that express prejudice without using humour. For this, the systems will be evaluated and ranked employing the F1-measure over the positive class.

2A-Prejudice Target Detection:

Taking into account the minority groups analyzed, i.e, Women and feminists, LGBTIQ community and Immigrants, racially discriminated people, and overweight people,  participants are asked to identify the targeted groups on each tweet as a multilabel classification task. The metric employed for the second task will be Weighted-F1.

2B-Degree of Prejudice Prediction:

The third subtask consists of predicting on a continuous scale from 1 to 5 to evaluate how prejudicial the message is on average among minority groups. We will evaluate the submitted predictions employing the Root Mean Squared Error.