PsExAI: Explainable AI for Clinical Psychology
Developing explainable artificial intelligence methods for clinical psychology applications.
PsExAI is a research project funded by the University of Bern DigiK Perspektivprojekte initiative (2024–2026). It aims to bridge the gap between high-performance machine learning models and their interpretability in clinical psychology settings.
Objectives
- Develop and evaluate explainability methods tailored to clinical psychological data and models.
- Create tools that help clinicians understand why a model makes a specific prediction, fostering trust and adoption.
- Investigate causal reasoning approaches to move beyond correlational pattern recognition.
Motivation
Machine learning models in mental health often operate as “black boxes.” While they can achieve strong predictive accuracy, clinicians and patients need to understand the reasoning behind predictions — especially when those predictions inform treatment decisions. PsExAI addresses this challenge by combining state-of-the-art explainable AI techniques with domain-specific clinical knowledge.