ALiDeR

Multimodal Speech-Based Depression Detection

Audio and text-based modelling for detecting depression-related patterns from clinical interview speech.

FedMPA

Federated Multimodal Learning

Privacy-performance optimisation for healthcare AI where sensitive data cannot be centrally shared.

PPS

Privacy-Performance Trade-offs Analysis

Quantifying the trade-off between predictive performance and privacy preservation to support trustworthy federated AI for healthcare.

Interpretable AI

Representation-Centric Interpretability

Moving beyond prediction-only explanations to inspect model representations and clinical alignment.

Clinical AI

Responsible Translation

Exploring safety, limitations, privacy, and clinical usefulness of AI-based mental health screening.

Generalizability

Beyond Depression

Extending the applicability of multimodal AI models to other mental health conditions and populations.