2010年7月16日星期五

Semantic Web is how it works?

In order to achieve the Semantic Web Intelligence and automated processing of information, goals, Semantic Web researchers have developed many new technologies and proposed a series of standards and norms. Widely recognized framework for the Semantic Web system by Tim Berners-Lee put forward, as shown in Figure 1, bottom-up a total of 7 layers.
W3C Semantic Stack Figure 1 Framework of the Semantic Web

Layer 1: URI and Unicode, the Semantic Web infrastructure.
Must first say a few words URI (Uniform Resource Identifier), the Chinese are usually referred to as "Uniform Resource Identifier", which is the foundation of the Semantic Web. Now identified with the use of the Internet URL, like HTML pages, Semantic Web also requires a similar specification, is used to uniquely identify the network resources. Semantic Web use URI specification, the specific specification in RFC 2396 for a detailed description. Any organizations and individuals are free to define and use the URI. Need to clarify a concept, URI and URL to use is very different. URL is used to identify a network path, can be accessed on the Internet at this URL to the corresponding resources; but the resource identified by a URI may be impossible to access to the network, it identifies and only identifies a resource, does not contain both The resources of the access path. Conceptually speaking, URI includes URL.
Surely you are not familiar with Unicode, it is a character set, the character set with all the characters are two bytes that can represent 65,356 characters, and basically includes all the world's languages characters. Semantic Web's ultimate goal is to build a global information network, it must cover all countries and all nationalities of the language, using Unicode as its character encoding scheme that can fundamentally solve the cross-regional, cross-language character encoding format standards.
This layer is the basis for the Semantic Web, URI for resource identification, Unicode encoding for resources.

Layer 2: XML, Namespaces, XML Schema and XML Query, Semantic Web syntax layer.
XML data representation and has become a de facto standard for data exchange, it provides a free format syntax, users can mark their own need to create the set, and write XML documents using these tags. Because anyone can freely define tag, so inevitably the situation will occur with the same name tag. W3C introducing Namespaces, which namespace mechanism, adding the tag URI index, and thus clears the conflict. For example, in which marked designation of the "cs" namespace to avoid conflict with other blog's tag. XML Schema provides an XML document on the mechanism for data validation. It is based on XML syntax, providing a variety of data types, the structure of the XML tags and use of standardized methods. XML XML Query is developed on the basis of technical standards, similar to XPath, etc. also use these techniques, data can be retrieved XML document, extract the node and other operations.
However, with XML in data exchange, application integration and other areas of popular, people gradually discovered, XML only describes the syntax of that data, but can not cover the semantics of the data. Given the general support of XML, the industry, and already has a more comprehensive technical standards, in the framework of the Semantic Web systems, as the syntax of data representation layer.

Layer 3: RDF Model, RDF Schema and RDF Syntax, Semantic Web data interoperability layer.
Not suitable for the expression of the semantics of XML data, therefore, the definition of the semantics of data and interoperability, which requires a higher level to complete. Therefore, W3C organization has developed a new language to describe Internet resources and their relationship, that is, resource description framework (Resource Description Framework, RDF). RDF uses triples to represent Internet resources, attributes and their values. In the Semantic Web concept, this is known as RDF triples the "statement (Statement)": subject, that is a resource (Resource), one of the property (Property) (verb) the value of an object (a resources, or the original value, such as the string).
RDF provides a standard set of data semantic specification, but it also need to define terms used in the description. RDF Schema (RDFS) provides a vocabulary for the computer to understand the definition of classes and properties provide the ability to describe. RDFS the RDF based on the introduction of the class, the relationship between class, property relations, and attributes such as domain and range. It is like a dictionary, the computer can understand by the meaning of the data. It has a significantly different from the object model, the property is independent of the class, an attribute can be applied to multiple classes or instances.
RDF Syntax constructed a complete set of syntax in order to facilitate automated computer analysis and processing. There are three common representations: graphic, N3, and XML. Among them, the graphical representation of the RDF model is a direct description of the model through the graphical RDF, RDF can understand the observed data and their relationships; N3 is a way to use triple by enumeration every RDF model "statement" to express the RDF model, it is the most easy to use, easy to understand; RDF / XML to RDF to describe XML grammar, the XML-parsing and interpretation process of integration, such, RDF parser to read XML in helping the same time, can get the XML to express the semantics and relationships according to their reasoning, and thus to make judgments based on semantics. However, RDF / XML is often because it is too complicated, difficult to use and be blamed.

Layer 4: Ontology, Semantic Web knowledge of the collection.
Ontology (Ontology) was first a philosophical concept, the scope of philosophy, the body is an objective explanation or description of a system, concerned about the abstract nature of objective reality. Artificial intelligence researchers first introduced to the computer field. On the body, one of the most popular definition is that "ontology is a shared conceptual model of the formal specification." This includes the meaning of 4 layers: conceptual model (conceptualization), clear (explicit), formal (formal) and sharing (share).
1) conceptual model (conceptualization): the objective world through abstract concepts in a number of phenomena which are the model, meaning that independent of their specific environmental conditions.
2) clear (explicit): the use of the concept and use of these concepts have a clear definition of the constraints.
3) formal (formal): Ontology is a computer-readable.
4) sharing (share): ontology is reflected in the common recognition of knowledge, and reflecting the related areas recognized set of concepts, it is aimed at groups rather than individuals consensus.
The goal of the extraction ontology domain knowledge, unified field of knowledge on a common understanding, identify areas of common recognition of the vocabulary, and from different levels of formal models of these terms and vocabulary given the interrelationship between clearly defined.
Semantic Web as the most core layer, ontology layer in the basic RDF and RDFS classes / attributes describing the basis of further describe the ontology and the relationships between them. This layer has its own specific ontology language. Historically there have been some impact, including, SHOE (Simple HTML Ontology Language), OIL (Ontology Inference Language), DAML (DARPA Agent Markup Language), and DAML + OIL. RDF is a simple ontology language, but its ability to describe the weak, the need for expansion. OWL (Web Ontology Language) is a W3C-recommended ontology language, it is more a reference implementation of the DAML + OIL design ideas and experience.

Level 5: Logic, rules and descriptions is the basis for automated reasoning.
Semantic Web is an important goal, is to realize automatic reasoning based on specific rules. This is a very complex task. Initially, around how to achieve this goal, or even whether we can achieve this goal, the researchers heated debate between. A typical topic is "RDF whether they have the ability to automatically reasoning?" Some people think that RDF / RDFS to quantify the lack of computing capacity.
For example, if your friend says, "" A Dry Story "is a good movie," You just read "A Story dry" and with his agreement. In this way, you will have a good friend of the judge to identify with. This recognition will certainly play a role in your reasoning. When he once again told us, "" The Promise "bored", you can based on the results between the judge on this comment.
Real life, people can make their own personal experience reasoning. In the Semantic Web, the researchers need to use this reasoning and quantitative methods to achieve that. In recent years, with the deepening of research, description logic (Description Logic, DL), as a more a more mature knowledge representation, to be entered into. OWL and other specifications of its designation, played a guiding role to some extent. Recently, researchers have begun to try to add rules in the formation of OWL OWL rules language ORL (OWL Rules Language), which can be better automated reasoning work.

6 layer: Proof, reasoning should be able to verify the results.
At present, the Trust for the Proof layer and the layer behind the research is not a lot, but a general consensus that, Proof and Trust within the field of the Semantic Web are two very important research topic. Proof Layer Logic layer using the definition of logical inference rules, draw some conclusions. For the Semantic Web users speaking, this reasoning should be based on reliable data based on the reasoning process should be open, and reasoning to the conclusion it should be verifiable.
Also use the example above, we reasoned that "" The Promise "is not a good movie." In real life, we can personally see "The Promise", and made judgments and conclusions on this reasoning to verify. In the Semantic Web, we must have such a verification mechanism. Computer concept, usually referred to as similar to the personal experience of "context." Context is made the basis of reasoning, therefore, the Semantic Web must establish a method of reasoning used to prove the accuracy and authenticity of context information.

Layer 7: Trust, Semantic Web should be a trusted network.
Reasoning within the Semantic Web and eventually come to the conclusion should be able to trust. This requires to satisfy two things: (1) can trust the data seen that context; (2) can trust the reasoning process. Meet these two before we can trust the final result by reasoning.
RDF model using the Semantic Web, anyone can describe any resources, the different positions of the people make the same resources may be the opposite of the description. Trust is responsible for the application layer to provide a mechanism to decide whether to trust the given proof. The establishment of Trust level, so that intelligent agents on the network to achieve personalized service, and automatic interaction between co-operation with the reliability and security.

Other: Signature, Digital Signature.
Hierarchical model of the digital signature in the right, runs through the middle of four semantic web. Digital signature is an Internet-based security authentication mechanisms. When information passes from one level to another level, you can use a digital signature description of information sources and security; this way, the receiver can identify its source by a digital signature and security, to determine the extent of information trusted. With digital signatures, a number of important e-commerce activities can be safely carried out in the semantic web. In fact, not only for the Semantic Web, the digital signature for all of the information exchange system are very important.

没有评论:

发表评论