Assessing similarity among data items becomes harder and harder, as the amount of semantics that it is daily added to the Web is continuously increasing. In terms of machine understanding, processing and interpreting this information requires novel fast and scalable data representations.
In the geospatial context, the geographical component is no more enough to assess connections between elements. In particular, geographical objects can be aggregated by position, category, density, and many other dimensions.
Embeddings allow representing the content of geographical objects in terms of vectors in a high dimensional space. In such a space, the distance reflects the “semantic distance” that holds among objects. This novel representation opens the door to the integration of geospatial ontologies into machine learning algorithms. The aim of this project is to investigate the coherence between geographical objects and embeddings, with specific attention to real applications (e.g., microblogs geolocation). Along with applications, we work on the theoretical side to deeply understand embeddings and proposing alternative algorithms to embed data.