I’m currently interested in what people actually mean when they call a dataset “diverse”. I work on embedding-based measures of data diversity: what they capture, how they differ, and which of them hold up.

The work is part of the DataDivers project at Utrecht University.