Semantic Web

While the Internet is a huge mass of information, its primary representation is HTML and is useful only to humans . The Semantic Web was created to change this by making the information useful also to machines. This process will not only make Web searches more specific, but will permit the representation of Internet information as knowledge that can then be tapped on a global scale.

The Semantic Web is still in its infancy, but is growing quickly since it will be a very important development. It was created by Tim Berners-Lee, the originator of the World Wide Web and the protocol on which it exists, HTTP. The Semantic Web is built upon two basic technologies, also relatively new. The first is XML (eXtensible Markup Language), which is an encoding scheme that permits wrapping information with tags that identify what the information represents. For example, if a Web page included the following information:

 <BR> Part Number: 2N2222 <BR> Type: Transistor <BR> Leads: 3 <BR> Cost: 
 <BR> Part Number: 2N2222 <BR> Type: Transistor <BR> Leads: 3 <BR> Cost: $0.25 
 .25 

This could be represented in XML as:

 <discrete-part>       <part-number>2N2222</part-number>       <type>Transistor</part-type>       <lead-count>3</part-lead-count>       <price-usd>0.25<price-usd> </discrete-part> 

In this case, there's no ambiguity about what the information represents. The information is easily parsed from the XML using the tags as markers for what is being represented. From the initial HTML representation, if a manufacturer used the term "Pins" instead of "Leads," then the meaning would break down. Additionally, in HTML, is the price U.S. dollars or Australian dollars? The XML tag makes this very unambiguous. Therefore, from this simple example it's clear that having an information schema can lead to more globally useful information. XML provides the ability to build up schemas so that a transistor could be represented by its own set of tags, which further simplifies this process.

The second technology that provides the basis for the Semantic Web is RDF, or Resource Description Framework. RDF is an XML-based language that permits information on the Web to be machine-processable, unambiguous, and meaningful. RDF is coded in triplets where each element of the triplet is an URI, or Uniform Resource Identifier. A URL (Uniform Resource Locator), of which a link on a Web page is the most common use, is a URI. The elements in an RDF triplet can take the form subject, verb, and object. Therefore, we can tie information together on the Web to draw similarities or associations.

Recall from the discussion of XML that schemas can be defined so that terms have a common meaning. With RDF, a corresponding ontology is used. An ontology is a document (referenced by RDF triplets) that defines the classes of objects and relations among them. Ontologies provide the mechanism for understanding data in a particular domain, in addition to drawing inferences about the data. This allows new knowledge to be created about existing knowledge within the Semantic Web. There are many challenges in this space, but the Semantic Web could be one of the most useful developments for AI.