Description
Simon Scheider will talk about Transformative geographic question answeringThe recent trend of adopting Artificial Intelligence (AI)-based methods has spurred hopes that geographic information can be successfully reused across disciplines without requiring the technical skills of GIS. Geographic question-answering (GeoQA), answering questions with geographic information, has therefore become an area of interest. However, in such systems, answers are usually directly matched to questions. Yet, most questions data scientists want to answer require some form of geo-analytical transformation of maps. Answers need to be given indirectly, by regenerating maps from various geodata sources, rather than retrieved from storage. Regarding this problem, I will report on the insights gained in a recent European Research Council (ERC) funded project (QuAnGIS). Furthermore, I will give an outlook on transformative GeoQA, wich is a way to interpret geographic questions in terms of the procedures for answering them. This may allow describing geodata as a source or purpose of geo-analytical transformations.
| Period | 2025 |
|---|---|
| Event title | Geospatial meets AI: The power of a self-conscious location |
| Event type | Seminar |
| Degree of Recognition | International |