You are invited to participate in the first CL-Aff Shared Task, to be held as a part of the Affective Content Analysis workshop @ AAAI 2019. In the quest to understand user expression, we propose a task focusing on one facet of human affect - happiness. We contribute a new labeled corpus of happy moments and pose two novel challenges to spur the development of supervised and semi-supervised approaches to model human affect.

The purpose of the CL-Aff Shared Task is to challenge the current understanding of emotion analysis through a task that models the experiential, contextual and agentic attributes of happy moments. It has long been known that human affect is context-driven, and that labeled datasets should account for these factors in generating predictive models of affect. The Shared Task is organized in collaboration with researchers at Megagon Labs and builds upon the HappyDB dataset, comprising human accounts of 'happy moments'. The Shared Task comprises two sub-tasks for analyzing happiness and wellbeing in written language, on a corpus of 100,000 descriptions of happy moments.

Tasks

GIVEN: An account of a happy moment, marked with individual's demographics, recollection time and relevant labels.

TASK 1: WHAT ARE THE INGREDIENTS FOR HAPPINESS ? 

Semi-supervised learning task: Predict agency and social labels for happy moments in the test set, based on a small labeled and large unlabeled training data.  

TASK 2: HOW CAN WE MODEL HAPPINESS ? 

Unsupervised task: Propose new characterizations and insights (not necessarily and not limited to themes) for happy moments in the test set, e.g., in terms of affect, emotion, participants and content.

Corpus details

• Labeled training set: Single-sentence happy moments from the available HappyDB corpus, annotated with labels that identify the 'agency' of the author and the 'social' characteristic of the moment, as well as concept labels describing its theme
• Unlabeled training set: The remaining single-sentence happy moments with no labels.
• Test set: Previously unreleased, labeled, single-sentence happy moments, freshly collected in the same manner as the original HappyDB data

Check out our brand new Git repository!

Updated FAQ !! - See https://github.com/kj2013/claff-happydb/blob/master/FAQ.txt

Upcoming deadlines

Shared Task Submission Deadlines: Submissions via EasyChair (prefix your submission title with "[CL-Aff Shared Task]")

• Aug 25, 2018 - Training set released
• October 15, 2018 - Abstract submission with a short system description
• October 30, 2018 - Test set released
• November 30, 2018 - System runs due & AAAI conference early registration deadline
• December 5, 2018 - System reports (full papers) due and peer review process begins
• TBD - Camera-ready contributions

January 27/28: All participants present at the Workshop at AAAI-19

Questions?

Contact Kokil Jaidka (jaidka at sas.upenn.edu) or Niyati Chhaya (nchhaya at adobe.com) and we promise to help.