Spatial Knowledge and Artificial Intelligence Lab (SKAI Lab)
About SKAI Lab
Torsten Hahmann’s laboratory research resolves around how to construct and integrate rich semantic descriptions of the information from complex and heterogeneous information systems in formal logical representations (called ontologies or knowledge graphs). Key questions in this work are how to manage and break-down rich ontologies, how to verify ontologies, how to combine them (integration), and how to automate the verification and integration process. SKIA mostly involves work with ontologies that capture some kind of spatial knowledge, which play a key role in geographic information systems, CAD and CAM software, mapping applications, environmental data science, and which serve as a testbed for methodological advances. The work advances knowledge representation, artificial intelligence, data science, logic, and various domain sciences (e.g. geography, earth sciences, life science, environmental science, disaster management).
Active Research Areas
- Formal representations of space, in particular spatial ontologies, and reasoning with them
- Qualitative representations of space, in particular mereotopology, incidence geometries, and betweenness relations
- Multidimensional representations: integrations of two-, three-, and four-dimensional spatial information
- Spatial intelligence that combines high-level qualitative conceptualizations with low-level geometric spatial information
- Commonsensical and scientific representations of physical space: materials, granularity
- Application-specific spatial ontologies: earth science (geological, hydrological, environmental) data, urban planning data, transportation data, building information, product specifications
- Automated methods and tools for ontology modularization, verification, comparison and integration relying, for example, on first-order theorem provers.
- Paradigms to combine different kinds of representations
- Integration of lightweight and expressive ontologies
- Integration of high-level knowledge with low-level data
- Integration of qualitative and quantitative knowledge
Torsten Hahmann Co-PI in developing The Urban Flooding Open Knowledge Network
NSF has awarded a multi-disciplinary research team involving the University of Cinncinati, North Carolina State University, the University of Illinois at Urbana-Champaign, Purdue University, and the University of Maine a $1,000,000 grant to develop “The Urban Flooding Open Knowledge Network” to improve prediction, risk analysis and mitigation of flooding in urban areas. The open knowledge network will connect public data and knowledge about the natural environment, such as from weather forecasts or terrain and elevation maps, with information about the water, sewer and energy infrastructure that are often affected during severe weather events. Read more about this award.