Shared-Task 1: Empathy Detection and Emotion Classification
We add to the previous shared task a new track on empathy, emotion and self-disclosure detection in conversation at the speech-turn level, using a new, unpublished dataset. This dataset is a multi-level of contextualized conversation that includes participants’ traits, their self-reported empathetic response to news articles, their perception of their conversational partners’ empathy, and turn-by-turn third-party assessments of the level of self-disclosure, emotion, and empathy expressed.
Shared-task 1 can be found on codalab.
Shared-Task 2: Emotion Classification on code-mixed text messages (Roman Urdu + English)
A novel shared task will be introduced focusing on multi-class and multi-label emotion classification on code-mixed (Roman Urdu + English) text messages. For this purpose, a dataset of 11,914 SMS messages will be used, which have been manually annotated with 11 emotions and the neutral class.
Shared-task 2 can be found on codalab.