Blueprint for the new generation of Web-semantic Web

With the increasing requirements on the use of network information, people have put forward higher requirements on the Content of network information. We can think that the network only provides people with an information sharing and information browsing environment. People can find the information they want in the network environment, people need to understand the information. That is to say, the information on the Internet is what people can understand, not what machines can understand.
Although massive amounts of information are distributed on the Internet, they are mainly oriented to humans. Because the information content is not better formally expressed, it is difficult for computers to process the information. On the Internet, the heterogeneous information formats, the diversity of information semantics, and information are widely used. LinkLack and non-uniformity, causing great inconvenience to people in information search, extraction, expression, interpretation and maintenance. Because of this, it is very difficult to carry out in-depth network applications, such as e-commerce, e-government, Digital Library and other intelligent services. In addition, because computers have the ability to process large-scale information, it is critical to automatically process and utilize online information to computers as much as possible. To achieve this goal, people must enable computers to "understand" the information and make better use of it on the premise of "Understanding. Tim Berners-Lee, the inventor of WWW, describes the blueprint for building the next generation of Web-the Semantic Web (Semantic Web) based on the existing Web ).
Semantic Web language tower andOntology
1. OntologyTheory
Semantic Web uses a certain Knowledge Representation Method to grant certain semantics to webpage information, so as to exchange and process semantic information between humans and computers. In such a huge information resource as Web, the knowledge base is diverse, including both common knowledge bases and domain knowledge bases in various fields. To ensure that the information in the network environment can be understood and interacted by computers, a unified presentation language is required for semantic-based unified presentation and interaction of Web Information.
OntologyThis topic provides a unified set of terms for members in the same application field and can conceptual representation of the descriptive objects. One OntologyConcept class, LinkFunctions, functions, and instances. OntologyThe concept is a broad concept. It can be a general concept, a task, a function, a behavior, a strategy, a reasoning process, and so on. OntologyThese concepts usually constitute a classification level. OntologyIn LinkIndicates the association between concepts. This Association represents the division of classification levels. LinkAll the relationships between concepts other than Functions Link, And LinkDifferent Support is between concepts LinkOr corresponding concepts. For example, the "ball volume" defines that the ball volume is uniquely determined by the ball's radius. In many fields, there are also associations or constraints between functions or associations. An instance is a basic element of a basic concept class. It refers to a specific entity and all instances in a specific field.
Semantic Web OntologyRepresents a unified and essential understanding of concepts in a specific field. For network applications, it is important to define a language with unified syntax, so that OntologyAbility to exchange information according to the unified syntax format.
2. Language tower in semantic Web
Starting from the current status of Web technology, Berners-Lee and others proposed a semantic Web layered architecture similar to the network model, which laid the foundation for the development of Semantic Web (1 ).
Figure 1? Semantic Web language Tower
At present, in the Semantic Web language Tower, the first two layers (Unicode & URI layer and XML + NS + xml schema layer) have become international standards for semantics and are widely used. The third-level RDF + rdfs has become the Recommendation Standard. OntologyLayer description languages include DAML+ OIL, and the standard OWL recommended by W3C, developed on this basis.
Study on language standards of different levels in semantic Web
OntologyIt provides a unified set of terms for members in the same application field. These members are people or smart agents. In this way, you need to develop a Representation Language to define the descriptive object. The existing expression languages and systems can be divided into two types: one is the expression method based on the first-order predicate logic, they are OntologyLanguage (KIF-based Ontololingua), Loom, and Frame-Logic. These methods have different expressions and computing features. However, for applications on the Internet, it is more important to define a language with unified syntax so that OntologyInformation exchange follows a unified syntax format. XML has become a standard language for data exchange on the Internet. It has good scalability, separation of style and content, and strict syntax requirements that can be followed, supports information exchange between different systems. Therefore, in order to simplify the analysis tasks of the Language Analyzer, we hope to establish XML-based syntax and support OntologyThe language standard of information exchange leads to the second type of XML-based OntologyLanguage Standards. These languages include SHOE (Simple HTML Ontology Extensions), XOL (Ontology Exchange Language), OML (Ontology Markup Language), RDF (Resource Description Frame ),??? OIL (Ontology Interchange Language ), DAML+ OIL (DARPA Agent Markup Language + Ontology Inference Layer) and OWL (the Web Ontology Language ). Although these languages are based on XML, they have different layers.
Based on OntologyBasic Semantic Web software (Semantic Web Infrastructure)
1. Early versions based on OntologyBasic Software Research
OntoKnowledge.org is a research project composed of multiple European research institutions and universities such as BT and Swiss Life. OIL (Ontology Inference Layer) OntologyThe language provides a hierarchical view that has become DAML. These organizations have also developed many research tools, the most mature of which is Oiled, which is OntologyEditing and design tools, supported DAMLFormat.
SiRPAC (Simple RDF Parser & Compiler) is an RDF interpreter and is widely used in many semantic Web projects developed by Stanford University. SiRPAC is actually an rdf api that can be accessed and stored on the basis of a file system Ontology. Sesame is a development source code OntologyStorage and Query System, developed by Aidministrator.
In addition, Chimeara of the KSL lab at Stanford University is currently applicable to OntologyOne of the most mature results of the study.
2. Study of Karlsruhe University
KAON (The Karlsruhe Ontology and Semantic Web Infrastructure) is a scientific research project at Karlsruhe University in Germany. This project is committed to providing the necessary foundation for semantic Web OntologySystem and related tools. It targets OntologyThe demand for upper-level commercial applications provides an open OntologyManagement Software, OntologyProvides a comprehensive support platform for storage, creation, and identification.
Figure 2? Architecture of the KAON Platform
The rdf api uses the rdf api of Stanford University, but it is rewritten and extended to provide upper-layer applications or KAON APIs. OntologyMemory storage mechanism. Currently, the rdf api includes not only an RDF Parser that can parse the RDF file, but also an RDF Serializer that can OntologySerialize LinkDatabases and files.
Kaon api shields applications from the underlying storage mechanism, but in fact it can also access KAON in multiple ways Ontology, Either through the rdf api (and then through the RDF Server), or directly through the Engineering Server. The definition of kaon api has its rationality. For example, it has the Observable design paradigm, which can be automatically obtained by applications. OntologyModify or upgrade the message.
Both the RDF Server and Engineering Server are based on LinkType database, which can provide concurrency control and transaction mechanisms. They can also directly support EJB (optional) and provide Entity Java Beans interfaces. The difference is that the RDF Server is oriented to RDF, and the Engineering Server is oriented to KAON. OntologyStandard. Another feature of Engineering Server is that it has a Preliminary distribution mechanism with clients.
The RDF Crawler of KAON is used for crawling and integrates the RDF information on the Web. You can put parameters such as the depth and specified range of the crawler into the configuration file and save the results to the local file. The KAON Portal is used to create a multilingual Ontology. You must first OntologyId. On the website OntologyFor visualized browsing and navigation. It strictly separates the display from the content. Good configurability. The OI-Modeler of KAON is OntologyModeling tool used to visually create a file and maintain it.
3. HP's Jena OntologyProcessing Platform
Jena at the HP lab is also a tool that uses the rdf api as the core to implement the Semantic Web (Sematic Web. Its architecture 3 is shown in.
Figure 3? Jena Architecture
As shown in figure 3, Jena consists of several parts, including Network API, Query, Readers, Inference, Writers, and Stores, which are centered on the rdf api.
4. Web-based distributed OntologySystem
The WODOS (Web Oriented Distributed Ontology System) System was developed by the Knowledge Engineering Office of the computer department of Tsinghua University. It is a comprehensive OntologySemantic Web Knowledge Processing Platform, supporting both Chinese and English. OntologyInformation stored in LinkIn a database, you can import and export RDF files and support RQL queries. Third-party software can perform operations through OdoAPI. WODOS introduces a reasoning mechanism to store data in WODOS OntologyFacts apply rules (based on Flogic) and reasoning based on these rules. WODOS establishes a unified user interface to integrate various WODOS operations (such as the original RQL query, the import and export of RDF files, and the newly added Flogic operations.
Web InformationOntologySemantic Content Conversion
OntologyGrant Semantic Web semantic information. OntologyThe operation of Internet information has become an important field of Semantic Web research. Mainly reflected in:
1. OntologyAnnotation
Add the information and OntologyLibrary OntologyAssociation is the primary task of Semantic Web development. Semantic Web annotation is to establish this association, generally using RDF or DAMLMark the Web. Web tagging was first started in Mosaic1.2 10 years ago. Since then, Web tagging tools have been developed, such as CritLink and ThirdVoice.
Many Web tagging systems were developed in subsequent studies, including CritLink, ThirdVoice, IMarkup, XLink, CREAM, and Annotea. However, all these tagging systems require manual operations to achieve tagging, that is, you can manually select the information to be labeled in the Web document and the corresponding OntologyTo enable Web tagging. This annotation method is ideal when marking a small number of pages, but this manual method is not applicable if you need to mark a large number of documents. Therefore, research on semi-automatic labeling has begun, and S-CREAM is a semi-automatic labeling system under study, which uses Amilcare as an Information Extraction Tool. Amilcare can be used as a training set through a set of manually labeled Web documents. Then, the learning algorithm is used to automatically extract the Extraction Rules. Finally, the rules are used to automatically mark other similar documents. S-CREAM can achieve semi-automatic annotation, but the automatically labeled document must be the same as the original document, such as the page that describes personal information.
2. OntologyIntegration
Because semantic Web works in a very scattered way, in the same field OntologyIt will be edited, created, and released in multiple places. OntologyAssistance is needed, and there may be semantic heterogeneity between them. How to solve this semantic heterogeneous problem in different fields OntologyProviding a unified semantic access layer is currently a hot topic in semantic Web research. This requires different Ontology. Currently OntologyThere are two commonly used methods in ing: one is to use an existing intermediate OntologyFor different Ontology. In the middle of this approach OntologyIn general, it is open and standard. The other is to directly OntologyIng. Two Fields OntologyThere are many similarities, so we can OntologyIng. Two OntologyMap OntologyConcepts and LinkAfter the concept ing is completed, LinkThe ing is complete.
3. OntologyCreate
OntologyThe description language provides a description. OntologyStandard language, while OntologyIt is a common understanding of semantics in a specific field. Different fields have different fields. Ontology. A domain contains rich knowledge and vocabulary. How to define a domain? OntologySemantic Web has become an important research issue. Currently OntologyThere are two types of research, one is in OntologyThe creation tool supports completely manual domain creation Ontology. The other is to use a large amount of information that is currently contained in the text, use machine learning methods, and use clustering and other technologies to establish semi-automatic Ontology.
Application of Semantic Web
1. Application of semantic Web in Web Services
The combination of Semantic Web and Web Services is considered an excellent application of semantic Web. As the latest development result of Web technology, Web Services will change the existing Web application model in its appearance and promotion. However, to enable servers distributed on the Internet to interact more automatically and intelligently through the Web, it is necessary to solve the heterogeneous information formats, the diversity of information semantics, and information on the Web. LinkLack and non-uniformity.
Set OntologyThe introduction of the concepts and technologies of Web Services will fundamentally solve the above problems. In addition OntologyRich semantics and extensive LinkIt will transform existing Web Services into Intelligent Web Services, and make the transformation of Web implementation from automation to intelligence possible.
In OntologyThe concept and technology OntologyWeb service description, service assembly and decomposition, based on OntologyService issue and service management.
2. Application in E-commerce
Web technology has gradually penetrated into all aspects of our daily life, and e-commerce has become an important business model. Open and flexible customers LinkEstablishing and managing business models is an important factor in realizing an effective e-commerce system. This openness and flexibility can be achieved through OntologyThe technology has been solved. For example OntologyProduct Description and classification system, as well as the structured, standardized, and personalized processing of the business processing process has become a hot research topic in the field of e-commerce.
Conclusion
The introduction of Semantic Web has attracted wide attention from academia and has made great progress in just a few years. I believe that with the continuous efforts of academia and industry, Semantic Web will gradually implement a new Web that people imagine!