kno.e.sis - Projects

It has been said that "an object by itself is intensely uninteresting." To fundamentally understand any entity, you must examine how it relates to other entities in its world. In other words, relationships are what define semantics (e.g., "relationships at the heart of semantics"). Correspondingly, metadata models defined as part of the W3Cs Semantic Web Activity treat relationships as first class objects, and researchers in this community have made progress toward mechanisms for querying complex relationships between resources where an ontology provides the context or domain semantics. As part of a related project, SemDis: Discovering Complex Relationships in the Semantic Web, we introduced the concept of semantic associations as complex relationships between resources and defined a set of query operators, p, for querying semantic associations. Semantic associations are defined in terms of connectivity and similarity in RDF graphs. The majority of work in this area has focused almost entirely on thematic relationships between resources, for example the fact that two people deposited money into the same bank account or that two glycopeptides participated in the same biological process.

While thematic metadata can tell us much about how two entities are related, in many domains and applications, we cannot ignore how the entities are related in space and time. The GIS community has put significant effort into the ontological modeling of geospatial relationships and geographic entities and the use of ontologies for search and analysis of geospatial data. However, the power of information systems which integrate ontologies describing thematic aspects of entities with ontologies describing the geospatial and temporal world in which they interact has yet to be fully realized. An information system which captures all of these aspects has enormous potential in many application areas, such as national security, emergency response, and e-learning.

This project will develop a framework which allows for discovery and analysis of relationships in all three dimensions of information: thematic, spatial and temporal. To realize an extension of thematic analytics to spatial and temporal analytics, we envision a system that combines vast amounts of thematic metadata captured using semantic web data representations with digital geospatial data collected by the GIS community. In addition, we want to incorporate temporal metadata which will allow us to perform computations and analysis on the temporal properties of these relationships.

From a scientific perspective, we face many challenges in this project. We must research how to formally model spatial and temporal properties of entities and their relationships in the context of Semantic Web data models, and correspondingly, we need to identify and formalize various relationship-based operators over this model. We will also require technology for extraction and integration of spatial and temporal data from web sources, and we must research the application of spatial and temporal reasoning to help deal with non-metric relationships and incompleteness of information.

Spatial and Temporal Semantic Analytics