Semantic Web & Information Brokering: Opportunities, Commercialization and Challenges

Editor's Notes

1. This is a more formal definition of an iscape. W e say “distributed” because the information to answer the request can lie in different sources.
2. In the context of digital earth , iscapes serve a very important role. We have developed a framework to specify iscapes declaratively. The primary usage of iscapes is meant by students. Iscapes serves as a ideal platform for students to lean about phenomena as the the requests are preformulated by the administrator and all students need to do is to select parameters and click on the request.
3. This is the specification framework for an iscape. An iscape can basically consist of 6 components as described below Ontologies serve as the shared vocabulary. Relationships serve as the smantic correlation layer. We could use simulation to demonstrate a concept graphically. Ontology Constraints are constraints that we can specify on the ontologies involved in the iscape. Iscapes can yield a lot of metadata. The presentation serves to filter the metadata. Finally one of the most important components is the student component where a student can configure parameters and learn interactively from the iscape.
4. This is an example ontology developed for the geographic domain. We shall get back to this topic later.
5. For e.g., x < y is a relationship that may hold between x and y. We have come across relations like…”equals”, “less_than”, “is_a”, etc
6. Most of these relations are not powerful enough to correlate complex entities in many common (and natural) domain like geography.
7. Now, lets take an example to see how we design the “affects” relationship. We see that different components of a volcano can affect different components of the environment. Put together, they can describe completely how volcanoes affect the environment. In this case, lets look at a few example components that affect others. Pyroclastic flows, if they flow across crops, will destroy them. So, we can say, if area of Pyroclastic flows intersect the area of crops, Pyroclastic flows destroy crops. Also, if the ash particles strewn from the volcano disperse into the atmosphere as tiny particles, their size can determine if they have a cooling effect on the atmosphere.
8. Let us put these sub-relations down into words. We see here that all these follow a specific pattern. [Function (something) operator Function (something else)]. If we generalize this, we can see that FN(x) op FN(y) where x and y are sub-components of A and B ontologies respectively. This schema can be used to define relationships in any domain and examples in 6-7 domains are shown in the thesis report.
9. Let us first take the case of comparison of locations. When we say location of volcano = location of the environment, we don’t expect to match the exact point of the volcano and the location specified. In general, the volcano’s effects would be felt around a certain area surrounding the volcano. We model this by including a tolerance level within which we match the location points. In this way, we perform a sort of imprecise of fuzzy match and helps us remove geo-spatial inconsistencies. This mapping technique is standardized by the use of enclosing functions and overloading the operator. We have developed mapping functions for the geographic domain and we need only to plug-in any function if we need other functionality.
10. This is an example of temporal matching. We can find out whether the given volcano had an affect on the environment on the given date. We know that a volcano’s effects like lava flows, etc would continue for a couple of days. We can assume this as tolerance. If the given date falls within this tolerance, we return a successful match. In the case of an earthquake, the time period is in the range of minutes.
11. .All iscapes and their components are specified using the Extendible Markup Language. Every iscape has an id , name and description. The ontologies involved and the name of the remaining components are then embedded in the iscape.
12. For example, if the iscape administrator wanted tp specify that the volcano was a stratovolcano, he could specify the name of the constraints within the constraint tags.
13. Relations have mapping conditions and value conditions. Mapping conditions are functions that you could apply on ontological terms , for example the area function equates the bounding coordinates of two ontologies. Value conditions denote configurable relationship parameters.
14. This component specifies the actual constraint. Here , we see that the iscape id and constraint name are the same as in the base iscape . This is then followed by the actual constraint specfication.
15. We can see that several metadata attributes can be included in the result presentation. The presentation layer is needed as we can then filter out the metadata returned by the system as result.
16. In this component , we encode what parameters can be configured in the iscape. Here we see that , the location of the environment ontology can be configured and the values that this parameter can take are Hawaii and Kileau.
17. 1. Operations are important for the ADEPT system as they lend themselves easily for changing parameters and viewing different results for every set of parameters which are entered by the user. 2. Geography instructors use a lot of simulation models to explain various concepts of geography to their students.
18. 1. A cellular automaton model of urban growth 2. Urbanization, agricultural intensification, resource extraction, and water resources development are examples of human-induced phenomena that have significant impact on people, economy and resources 3. Based on an understanding of the land use changes, it may be possible to understand the impacts associated with them and contribute to a productive national environmental sustainability
19. This screen shows the student interface to ADEPT. We can see that the ontologies, volcanoes and environment are used here as well as the ontology country. All ontological terms and iscapes along with configurable parameters are embedded in the same screen.