Augmented Intelligence for Clinical Detection Support of Eating Disorders
As part of the research direction Augmented Intelligence for Mental Health Diagnostics, this project investigates the potential of using innovative technologies from Natural Language Processing (NLP) for the diagnostics of eating disorders.
Factsheet
- Schools involved School of Engineering and Computer Science
- Institute(s) Institute for Data Applications and Security (IDAS)
- Research unit(s) IDAS / Applied Machine Intelligence
- Funding organisation Others
- Duration (planned) 01.10.2022 - 01.03.2024
- Head of project Prof. Dr. Mascha Kurpicz-Briki
- Project staff Ghofrane Merhbene
- Partner INVENTUS BERN - Stiftung
- Keywords mental health, psychiatry, psychology, artificial intelligence, augmented intelligence, natural language processing
Situation
Natural language processing (NLP) and machine learning (ML) technologies provide new possibilities for tools in clinical psychology. Patient assessment based on inventories (questionnaires) can have limitations, and manually evaluated interview transcripts and notes are time-consuming. Automated text analysis technologies can provide a handy tool to the clinical practitioners in diagnostics. Promising results have been obtained for burnout in previous work, where machine learning was used to find indication for burnout in free-text. Other work in the current state-of-the-art is dealing with depression or suicide risk assessment.
Course of action
However, few works exist for text classification with regard to eating disorders. So far, this has been mainly explored for the English language. To provide new tools for clinical usage in Switzerland, local languages need to be explored, which includes particular challenges such as different cultural and linguistic structures, or potential lack of sufficient training data. These points require additional research to overcome. Furthermore, in existing work, applications are mainly targeting users e.g., as smartphone applications for self-help, and not clinical professionals. These limitations of the state-of-the-art will be investigated in this project.
Looking ahead
In the context of augmented intelligence supporting human decision-makers with their daily work, in this project a solution supporting clinical professionals with the diagnosis based on free-text questions or interview transcripts will be targeted. We explore whether NLP/ML technologies can be applied to detect indication for different types of eating disorders in German and French free-texts, and develop a prototype of a tool - using a fully human-centered approach - to support clinical practitioners in the future.
