Transformer-based language models for mental health issues: A survey.

• To overview Transformer models used for text-based mental health issue identification. • To characterize the methods based on Transformers developed in the past few years. • To highlight opportunities and challenges of the methods in addressing mental issues. Early identification and prevention of mental health stresses and their outcomes has become of urgent importance worldwide. To this purpose, artificial intelligence provides a body of advanced computational tools that can effectively support decision-making clinical processes by modeling and analyzing the presence of a variety of mental health issues, particularly when these can be detected in text data. In this regard, Transformer-based language models (TLMs) have demonstrated exceptional efficacy in a number of NLP tasks also in the health domain. To the best of our knowledge, the use of TLMs for specifically addressing mental health issues has not been deeply investigated so far. In this paper, we aim to fill this gap in the literature by providing the first survey of methods using TLMs for text-based identification of mental health issues. [ABSTRACT FROM AUTHOR]