Personalised narratives through knowledge graphs

Ontologies enable integration of heterogenous data from multiple sources and in different modalities (e.g. video, text). This creates large knowledge graphs of linked data that underpin a range of information exploration tasks. However, when users are unfamiliar with the conceptual structures encoded in the graph, which is inevitable with large graphs and complex domain knowledge, the exploration experience can be frustrating and overwhelming, and will have little utility for the user’s task.

These challenges will be addressed in this PhD project. It will develop novel automated methods for generating personalised trajectories through knowledge graphs to support sensemaking tasks. Heterogeneous data and existing ontologies produced from current projects will provide the practical need and the experimental platform for this project. Examples include (a) exploration of ontologies and data resources on urban infrastructure to facilitate decision making, or (b) knowledge-driven analysis and exploration of video snippets to facilitate medical consultation training.