《高级Web技术》参考资料：语义Web A Semantic Web Primer.pdf_P31-P35

TLFeBOOK section 1.2.4). Tim Berners-Lee speaks of an"Oh yeah? button that would ask for an explanation Explanations will also be necessary for activities between agents. while some agents will be able to draw logical conclusions, others will only have the capability to validate proofs, that is, to check whether a claim made by another agent is substantiated. Here is a simple example. Suppose agent natural language, of course, but in a forma sage"You owe me $80"(not in to agent 2, representing a person. Then agent 2 might ask for an explanation, and agent 1 might respond with a sequence of the form Web log of a purchase over $80 Proof of delivery(for example, tracking number of UPS) Rule from the shop's terms and conditions purchase(X, Item)A price(Item, Price)A delivered(Item, X) →oues(X, Price Thus facts will typically be traced to some Web addresses(the trust of which will be verifiable by agents), and the rules may be a part of a shared com merce ontology or the policy of the online shop For logic to be useful on the Web it must be usable in conjunction with ther data, and it must be machine-processable as well. Therefore, there is ongoing work on representing logical knowledge and proofs in Web lan guages. Initial approaches work at the level of XMl, but in the future rules and proofs will need to be represented at the level of RDF and ontology lan guages, such as DAML+OIL and OWL 1.3.4 Agents Agents are pieces of software that work autonomously and proactively Con ceptually they evolved out of the concepts of object-oriented pr ogrammIn and component-based software development. A personal agent on the Semantic Web(figure 1. 2 )will receive some tasks and preferences from the person, seek information from Web sources, com- municate with other agents, compare information about user requirements and preferences, select certain choices, and give answers to the user. An example of such an agent is Michael's private agent in the physiotherapy example of section 1.2.4 TLFebooK

14 1 The Semantic Web Vision section 1.2.4). Tim Berners-Lee speaks of an “Oh yeah?” button that would ask for an explanation. Explanations will also be necessary for activities between agents. While some agents will be able to draw logical conclusions, others will only have the capability to validate proofs, that is, to check whether a claim made by another agent is substantiated. Here is a simple example. Suppose agent 1, representing an online shop, sends a message “You owe me $80” (not in natural language, of course, but in a formal, machine-processable language) to agent 2, representing a person. Then agent 2 might ask for an explanation, and agent 1 might respond with a sequence of the form Web log of a purchase over $80 Proof of delivery (for example, tracking number of UPS) Rule from the shop’s terms and conditions: purchase(X, Item) ∧ price(Item, P rice) ∧ delivered(Item, X) → owes(X, P rice) Thus facts will typically be traced to some Web addresses (the trust of which will be verifiable by agents), and the rules may be a part of a shared commerce ontology or the policy of the online shop. For logic to be useful on the Web it must be usable in conjunction with other data, and it must be machine-processable as well. Therefore, there is ongoing work on representing logical knowledge and proofs in Web languages. Initial approaches work at the level of XML, but in the future rules and proofs will need to be represented at the level of RDF and ontology languages, such as DAML+OIL and OWL. 1.3.4 Agents Agents are pieces of software that work autonomously and proactively. Conceptually they evolved out of the concepts of object-oriented programming and component-based software development. A personal agent on the Semantic Web (figure 1.2) will receive some tasks and preferences from the person, seek information from Web sources, communicate with other agents, compare information about user requirements and preferences, select certain choices, and give answers to the user. An example of such an agent is Michael’s private agent in the physiotherapy example of section 1.2.4. TLFeBOOK TLFeBOOK

TLFeBOOK mally the beliefs, desires, and intentions of agents, and to create and main- tain user models. However, these points are somewhat orthogonal to the Semantic Web technologies. Therefore they are not discussed further in this 1.3.5 The Semantic Web versus artificial Intelligence As we have said, most of the technologies needed for the realization of the Semantic Web build upon work in the area of artificial intelligence. Given that AI has a long history, not always commercially successful, one might worry that, in the worst case, the Semantic Web will repeat Al's errors: big promises that raise too high expectations, which turn out not to be fulfilled (at least not in the promised time frame) This worry is unjustified. The realization of the Semantic Web vision does not rely on human-level intelligence; in fact, as we have tried to explain, the challenges are approached in a different way. The full problem of AI is a deep scientific one, perhaps comparable to the central problems of physics (explain the physical world)or biology (explain the living world). So seen, the difficulties in achieving human-level Artificial Intelligence within ten or twenty years, as promised at some points in the past, should not have come as a surprise. But on the Semantic Web partial solutions will work. Even if an intelligent agent is not able to come to all conclusions that a human user might draw, the agent will still contribute to a Web much superior to the current Web. This brings us to another difference. If the ultimate goal of AI is to build an intel- ligent agent exhibiting human-level intelligence(and higher), the goal of the Semantic Web is to assist human users in their day-to-day online activities It is clear that the Semantic Web will make extensive use of current Al tech- ology and that advances in that technology will lead to a better Semantic Jeb. But there is no need to wait until AI reaches a higher level of achieve- ment; current AI technology is already sufficient to go a long way toward realizing the Semantic Web vision 1.4 A Layered Approach The development of the Semantic Web proceeds in steps, each step building a layer on top of another. The pragmatic justification for this approach is that it is easier to achieve consensus on small steps, whereas it is much harder to get everyone on board if too much is attempted. Usually there are sev- TLFebooK

16 1 The Semantic Web Vision mally the beliefs, desires, and intentions of agents, and to create and maintain user models. However, these points are somewhat orthogonal to the Semantic Web technologies. Therefore they are not discussed further in this book. 1.3.5 The Semantic Web versus Artificial Intelligence As we have said, most of the technologies needed for the realization of the Semantic Web build upon work in the area of artificial intelligence. Given that AI has a long history, not always commercially successful, one might worry that, in the worst case, the Semantic Web will repeat AI’s errors: big promises that raise too high expectations, which turn out not to be fulfilled (at least not in the promised time frame). This worry is unjustified. The realization of the Semantic Web vision does not rely on human-level intelligence; in fact, as we have tried to explain, the challenges are approached in a different way. The full problem of AI is a deep scientific one, perhaps comparable to the central problems of physics (explain the physical world) or biology (explain the living world). So seen, the difficulties in achieving human-level Artificial Intelligence within ten or twenty years, as promised at some points in the past, should not have come as a surprise. But on the Semantic Web partial solutions will work. Even if an intelligent agent is not able to come to all conclusions that a human user might draw, the agent will still contribute to a Web much superior to the current Web. This brings us to another difference. If the ultimate goal of AI is to build an intelligent agent exhibiting human-level intelligence (and higher), the goal of the Semantic Web is to assist human users in their day-to-day online activities. It is clear that the Semantic Web will make extensive use of current AI technology and that advances in that technology will lead to a better Semantic Web. But there is no need to wait until AI reaches a higher level of achievement; current AI technology is already sufficient to go a long way toward realizing the Semantic Web vision. 1.4 A Layered Approach The development of the Semantic Web proceeds in steps, each step building a layer on top of another. The pragmatic justification for this approach is that it is easier to achieve consensus on small steps, whereas it is much harder to get everyone on board if too much is attempted. Usually there are sevTLFeBOOK TLFeBOOK

TLFeBooK 17 eral research groups moving in different directions; this competition of ide is a major driving force for scientific progress. However, from an engineer- there is a need to standardize. So, if most researchers on certain issues and disagree on others, it makes sense to fix the points of agreement. This way, even if the more ambitious research efforts should fail, there will be at least partial positive outcomes Once a standard has been established, many more groups and companies will adopt it, instead of waiting to see which of the alternative research lines will be successful in the end. The nature of the Semantic Web is such that companies and single users must build tools, add content, and use that con- tent. We cannot wait until the full Semantic Web vision materializes-it may take another ten years for it to be realized to its full extent(as envisioned today, of course) Downward compatibility. Agents fully aware of a layer should also be ble to interpret and use information written at lower levels. For exam- ple, agents aware of the semantics of OWL can take full advantage of information written in rdf and rdf schema Upward partial understanding On the other hand, agents fully aware of a layer should take at least partial advantage of information at higher levels For example, an agent aware only of the RDF and RDF Schema semantics can interpret knowledge written in OWL Partly, by disregarding those elements that go beyond RDF and RDF Schema t. Figure 1.3 shows the"layer cake"of the Semantic Web(due to Tim Berners- Lee), which describes the main layers of the Semantic Web design and vision At the bottom we find XML, a language that lets one write structured Web documents with a user-defined vocabulary. XML is particularly suitable for RDF is a basic data model, like the entity-relationship model, for writing simple statements about Web objects (resources). The rdF data model does not rely on XMl, but RDF has an XML-based syntax. Therefore, in figure 1.3, it is located on top of the XML layer RDF Schema provides modeling primitives for organizing Web objects into hierarchies. Key primitives are classes and properties, subclass and subprop- erty relationships, and domain and range restrictions. RDF Schema is based RDF TLFeBOoK

1.4 A Layered Approach 17 eral research groups moving in different directions; this competition of ideas is a major driving force for scientific progress. However, from an engineering perspective there is a need to standardize. So, if most researchers agree on certain issues and disagree on others, it makes sense to fix the points of agreement. This way, even if the more ambitious research efforts should fail, there will be at least partial positive outcomes. Once a standard has been established, many more groups and companies will adopt it, instead of waiting to see which of the alternative research lines will be successful in the end. The nature of the Semantic Web is such that companies and single users must build tools, add content, and use that content. We cannot wait until the full Semantic Web vision materializes — it may take another ten years for it to be realized to its full extent (as envisioned today, of course). In building one layer of the Semantic Web on top of another, two principles should be followed: • Downward compatibility. Agents fully aware of a layer should also be able to interpret and use information written at lower levels. For example, agents aware of the semantics of OWL can take full advantage of information written in RDF and RDF Schema. • Upward partial understanding. On the other hand, agents fully aware of a layer should take at least partial advantage of information at higher levels. For example, an agent aware only of the RDF and RDF Schema semantics can interpret knowledge written in OWL partly, by disregarding those elements that go beyond RDF and RDF Schema. Figure 1.3 shows the “layer cake” of the Semantic Web (due to Tim BernersLee), which describes the main layers of the Semantic Web design and vision. At the bottom we find XML, a language that lets one write structured Web documents with a user-defined vocabulary. XML is particularly suitable for sending documents across the Web. RDF is a basic data model, like the entity-relationship model, for writing simple statements about Web objects (resources). The RDF data model does not rely on XML, but RDF has an XML-based syntax. Therefore, in figure 1.3, it is located on top of the XML layer. RDF Schema provides modeling primitives for organizing Web objects into hierarchies. Key primitives are classes and properties, subclass and subproperty relationships, and domain and range restrictions. RDF Schema is based on RDF. TLFeBOOK TLFeBOOK