Geosocial Artificial Intelligence

Our research group is pioneering a dual approach that integrates both “sensing” and “actuating” processes. This methodology not only focuses on analysing social environments but also actively shaping them through data-driven insights. We aim to create novel AI methodologies that provide deeper understanding of geospatial and social processes by analysing human-generated data. This holistic approach to research transcends traditional disciplinary boundaries and integrates concepts from various fields like geoinformatics, computer science, sociology, psychology, education research, and others. 

Key research areas will include: 

Multimodal Machine Learning: Investigating methods for analysing and integrating heterogeneous data sources simultaneously considering multiple modalities. 

Information Fusion: Creating novel algorithms for combining various data streams into a single information layer. 

Generative AI and Active Learning: Exploring generative models to create synthetic data, involving a human-in-the-loop approach. 

Contextual Enrichment: Designing methods to contextualise information layers within relevant spatial, temporal, semantic or emotional contexts, thus increasing their interpretability.