In recent years, there has been an increase in interest and explorations of the use of machine learning (ML) to assist in the diagnosis of mental health problems; and for improving access to, engagement with, and the outcomes of, therapeutic treatment. Amongst a wide range of diagnosable mental disorders, affective disorders – such as depression, bipolar and anxiety – are the most common. For these disorders, distortions and inconsistencies in a person’s emotional state (their mood) present the primary cause for disruptions to their life. Here, ML promises to offer new routes for improving the identification of health risk factors; the prediction of disease progression; and the development of personalised health interventions. Research to date has started to explore the identification of mental health problems through inferences about peoples’ behaviours on social media, online searches, or mobile phone app uses; as well as varied approaches to assess, or continuously monitor, a person’s mental health and related symptoms by measuring sleep, mood, stress or physical activity via audio, visual or physiological signal processing.
Despite great potential, the realization of effective ML-enabled applications for mental health remains a hugely challenging area for research and development. Among the very many challenges in this domain are the need for a stronger focus on real-world applications and user-centred design processes to aid the identification of real healthcare needs that can sensibly be supported through ML; and accordingly, careful choices in data collection and the design of reliable and fair algorithmic models. Especially in mental health, where data and ML-supported decisions can have far reaching personal, social and economic impact, we need to be very critical of what reasonable inferences can be drawn from specific data; design interfaces that help people to appropriately interpret system inferences; and ensure that, ultimately, humans remain in control over, and accountable for, important ML-informed decisions.
This workshop seeks to bring together an inter-disciplinary group of researchers and practitioners from academia and industry to discuss the unique opportunities and challenges for developing effective, ethical and trustworthy ML- approaches and interventions for the diagnosis and treatment of affective disorders. The specific focus will be on (but it is not limited to):
We invite submissions of short papers (2-4 pages) in ACII paper format. Submissions will be reviewed by members of the organising and program committee based on relevance to the workshop and potential for contributing to discussions. Workshop proceedings will be published by IEEE Xplore.
Please submit your paper to the "Machine Learning for the Diagnosis and Treatment of Affective Disorders" track via the easychair submission system.
Associate Professor in Computer Science and Statistics at Trinity College Dublin,
UX Director SilverCloud Health
Saeed Abdullah, Penn State University, US
Talayeh Aledavood, University of Helsinki, FI
Angel Enrique, Silvercloud Health, IRL
Nadia Bianchi-Berthouze, University College London, UK
Rafael Calvo, Imperial College London, UK
Afsaneh Doryab, Carnegie Mellon University, US
Jean Marcel Dos Reis Costa, Cornell University, US
Marzyeh Ghassemi, University of Toronto, CA
Martin Gjoreski, Jožef Stefan Institute, SVN
Mark Matthews, HealthRythms, US
Tristan Naumann, Microsoft Research Redmond, US
Temitayo Olugbade, University College London, UK
Pablo E. Paredes, Standford University, US
Koustuv Saha, Georgia Institute of Technology, US
Björn W. Schuller, University of Augsburg, GER
Greg Wadley, University of Melbourne, AUS
Steffen Walter, University of Ulm, GER
Stay tuned, details to follow soon!
Assistant Professor of Psychological Science
University of California, Irvine