Findings of the Shared Task on Detecting Signs of Depression from Social Media

Abstract

Social media is considered as a platform where users express themselves. The rise of social media as one of humanity's most important public communication platforms presents a potential prospect for early identification and management of mental illness. Depression is one such illness that can lead to a variety of emotional and physical problems. It is necessary to measure the level of depression from the social media text to treat them and to avoid the negative consequences. Detecting levels of depression is a challenging task since it involves the mindset of the people which can change periodically. The aim of the DepSign-LT-EDI@ACL-2022 shared task is to classify the social media text into three levels of depression namely "Not Depressed", "Moderately Depressed", and "Severely Depressed". This overview presents a description on the task, the data set, methodologies used and an analysis on the results of the submissions. The models that were submitted as a part of the shared task had used a variety of technologies from traditional machine learning algorithms to deep learning models. It could be observed from the result that the transformer based models have outperformed the other models. Among the 31 teams who had submitted their results for the shared task, the best macro F1-score of 0.583 was obtained using transformer based model.