Xem mẫu

TLFeBOOK 8 Conclusion and Outlook 8.1 How It All Fits Together At this time it may be instructive to look back at chapter 1, where the Seman-tic Web vision was described. In this book, we described the key Semantic Web technologies. Now we consider an automated bargaining scenario to see how all technologies discussed fit together. • Each bargaining party is represented by a software agent. We have not discussed agents in this book and refer readers to the extensive litera-ture. Often, agents are treated as black boxes, which solve all problems miraculously. We preferred to concentrate on the internals of agents, and refrained from discussing aspects of agent communication and collabora-tion. • The agents need to agree on the meaning of certain terms by committing to a shared ontology, e.g., written in OWL. • Case facts, offers, and decisions can be represented using RDF statements. These statements become really useful when linked to an ontology. • Information is exchanged between the agents in some XML-based (or RDF-based) language. • The agent negotiation strategies are described in a logical language. • An agent decides about the next course of action through inferring con-clusions from the negotiation strategy, case facts, and previous offers and counteroffers. TLFeBOOK TLFeBOOK 224 8 Conclusion and Outlook 8.2 Some Technical Questions 8.2.1 Web Ontology Language: Is Less More? Much of the effort in Semantic Web research has gone into developing an ap-propriate Web ontology language, resulting in OWL as the current standard. One key question is whether the ontology languages need to be very com-plex. While one can always think of cases that one might wish to model and that are beyond the expressive power of full first-order logic, the question remains whether these issues are important in practice. There are reasons to expect that most ontological knowledge will be of a rather simple nature, and that less expressive languages will be sufficient. The advantages of simple ontology languages are a more efficient reasoning support, a simpler language for tool vendors to support, and a more easily usable language. The latter may turn out to be of crucial importance for the success of the Semantic Web. OWL Lite is a step in the right direction. 8.2.2 Rules and Ontologies As we said in chapter 4, the current (advanced) Web ontology languages are based on description logics. On the other hand, it has been recognized that rules are an important and simple representation formalism with many applications. Currently there is ongoing work on combining both. We believe that a formalism that combines the full power of both descrip-tion logics and rules would be overkill. Apart from questions regarding the need for such rich languages, the research has revealed several complexity and computability barriers that are difficult to overcome. A sensible compromise approach may be to take RDFS and put rules on top, as an alternative to going down the path of description logics. There are no real technical problems with this approach. And it is not as restrictive as it looks, because many features of description logics (and thus OWL) are definable using rules. 8.3 Predicting the Future So, will the Semantic Web initiative succeed? While many people believe in it (and in fact are investing in it), the outcome is still open. As suggested at the beginning of this book, the question is not so much a technological but rather a practical one: Will we be able to demonstrate the usefulness of this TLFeBOOK TLFeBOOK 8.3 Predicting the Future 225 technology quickly and powerfully enough to create momentum (recreating something similar to the early stages of the World Wide Web)? Where will the ontologies come from? We already see the solutions to this potential bottleneck: some large ontologies are becoming de facto standards (WordNet, NCIBI’s cancer ontology), and many small ontologies are either hand-created by organizations (e.g., RosettaNet) or by machine through ma-chine learning techniques, natural language analysis, and borrowing from legacy resources (e.g., database schemas). Where will the semantic markup come from? It is clear that the bulk of the required large volumes of semantic markup will not be created by hand (unlike the start of the World Wide Web, which did happen through hand-coded HTML pages). Instead, analysis of documents through natural lan-guage techniques and borrowing from legacy sources (e.g., databases) will be prominent techniques here. Where will the tools come from? This is a potential bottleneck that is al-ready in the process of being resolved. A large variety of tools is already available for every aspect of the Semantic Web application life cycle (editors, storage, query and inference infrastructure, visualization, versioning tools). Currentlythesetoolsaremostlyintheacademicdomain, buttheyarequickly being taken up by the commercial sector, in particular, by highly innovative startups, both in the United States and in the European Union. How should one deal with a multitude of ontologies? This problem (known as the ontology mapping problem) is perhaps the hardest problem to be solved. Many approaches are being investigated (based on negotiating agents, ma-chine learning, or linguistic analysis), but the jury is still out on this one. Possibly the first success stories will not emerge in the open heterogeneous environment of the WWW but rather in intranets of large organizations. In such environments, central control may impose the use of standards and technologies, and possibly the first real success stories will emerge. Thus we believe that knowledge management for large organizations may be the most promising area to start. Other areas that will be quick to follow are so-called e-science: the use of the Semantic Web by scientists (just as the use by scientists was an important catalyst for the World Wide Web). It could well be that e-commerce, with all its associated problems of privacy, security, and trust, will only be a later application of the Semantic Web. All in all, we are optimistic about the future of the Semantic Web and hope that this book as a teaching resource will play its role in “bringing the Web to its full potential”. TLFeBOOK TLFeBOOK TLFeBOOK TLFeBOOK A Abstract OWL Syntax The XML syntax for OWL, as we have used it in chapter 4 is rather verbose, and hard to read. OWL also has an abstract syntax1, which is much easier to read. This appendix lists the abstract syntax for all the OWL code discussed in chapter 4. 4.2.2: Header Ontology( Annotation(rdfs:comment "An example OWL ontology") Annotation(rdfs:label "University Ontology") Annotation(owl:imports http://www.mydomain.org/persons) ) 4.2.3: Class Elements Class(associateProfessor partial academicStaffMember) Class(professor partial) DisjointClasses(associateProfessor assistantProfessor) DisjointClasses(professor associateProfessor) Class(faculty complete academicStaffMember) 1. Defined in TLFeBOOK ... - tailieumienphi.vn
nguon tai.lieu . vn