Mr. Angele, what are semantic technologies?
Angele: Semantic technologies help create, represent, and process knowledge and information. They offer an opportunity to make information understandable to computers so that they can interpret it.
How do these technologies work?
Angele: The most important components of semantic technologies are ontologies – models that represent the relationships of knowledge. Ontologies allow us to represent terms and their characteristics and to describe relationships to other terms. Ontologies are also used to display relationships of any complexity. A simple example is the display of the relationship between the power output of a motor and the performance of the braking system: “Brake performance Y belongs to engine performance X.” If knowledge is mapped in such a model, specialized software products called inference machines can process it and independently draw conclusions from it.
What’s the current status of research internationally?
Angele: Right now, research is concentrating on developing semantic technologies for the Semantic Web and for Semantic Web Services. The Semantic Web will represent information in a way that computers can understand. It enables the development of new services, such as automatic travel planning and booking by software agents that independently interpret offers and process travel data. To solve the tasks related to such a service, we are researching how computers can draw conclusions automatically from ontologies, how ontologies are created, and which languages can describe ontologies. Languages that describe ontologies are also an important topic in the World Wide Web Consortium (W3C) because the Internet will probably develop into a Semantic Web.
How we assign ontologies, how we annotate texts (provide them with metainformation), and how we describe or configure Web Services also play a role. Internationally, the University of Karlsruhe and renowned research centers, such as the Universities of Innsbruck, Manchester, Stanford, Madrid, and Amsterdam, are in the lead.
Do semantic technologies let you jump from processing data to processing knowledge?
Angele: Yes. Semantic applications “understand” information. “Understanding” presumes a common language that excludes conceptual and terminological confusion, ambiguity, and equivocation. And that’s just what semantic technologies can provide. In an ontology, relevant terms and their contexts are defined precisely for an application area. The ontology describes a generally recognized understanding of the application area, an understanding shared and used by all persons and applications.
Where do you see opportunity to apply these developments?
Angele: The integration of data from various sources and the mapping of complex relationships are the most important areas. We call this “semantic business integration.” For example, mapping two databases, each designed to meet a different goal, is extremely complex and can be simplified with the aid of semantic technologies. The combination of integrating contents and mapping complex relationships enables us to fulfill the increasingly complex demands on IT.
We already use our technology to support various core processes in companies. Right now, we’re focusing on development processes in the automobile industry and on knowledge processes in industries that rely to a great extent on consulting. The centrality of information in itself is becoming ever more significant. In the future, processes will be started, controlled, and measured on the basis of information. It’s no longer enough to know what information specific processes require. Instead, we need to know what information enables specific processes or is required for them. In the future, the supporting IT software must cover changing information needs flexibly and dynamically.
The project that we’re now working on with a large automobile manufacturer is a concrete example. We’re realizing a system to configure test cars. We must consider complex relationships between individual components. The engineers must consider a number of logical, geometric, and functional relationships: which transmission fits with which motor, which braking system is appropriate, and how to adjust the electrical system. A large portion of the required information is already present in computer-aided design (CAD) and product data management (PDM) systems. Here we need flexible structures that build upon existing systems.
Other companies have already used semantic technologies for some time. For example, Boeing models the complex relationships in aircraft. A help desk system that contains complex knowledge in an ontology answers questions such as, “May I lay a special type of carpet at this location in the aircraft?” The answer to this question not only requires knowledge of fire prevention and emergency evacuation procedures, but also the experience of the engineers.
Semantic technologies are also used for information retrieval and grid computing. Information retrieval involves an intelligent search for documents; semantic grid computing distributes CPU-intensive tasks to a large number of PCs. With semantic technologies, we separate such tasks into smaller subtasks and ultimately combine partial solutions into an overall solution.
What will semantic technologies be able to do in the future?
Angele: The Internet will increasingly develop in the direction of a Semantic Web. Someone who wants to participate in a conference will simply send out a software agent. The agent will work with the user’s date book to register the user, select appropriate workshops, and find a hotel and flight that meet the user’s preferences.
How will suppliers of business software, such as SAP, profit from these developments?
Angele: Hasso Plattner once said, “The software industry is building an alphabet, but hasn’t yet invented a common language.” He alluded to the need for a uniform, descriptive language to integrate heterogeneous designs. Semantic technologies offer this uniform description because they describe the meaning and context of information. Ontologies can also depict complex relationships between knowledge. We separate the logic and the meaning of the information from the data sources and applications that provide them, which enables us to create a flexible, serviceable, and, above all, reusable interface.
Where are the greatest difficulties in the further development of semantic technologies?
Angele: Text analysis is one of the greatest challenges. It is very difficult to have a computer automatically analyze unstructured textual information, as present in normal texts and conversation.