Semantic Web

Overview
The Semantic Web, as termed by Tim Berners-Lee, is an approach that promotes standards when it co

mes to data formatting, allowing data to be shared across various web platforms. Essentially, it is an extension of the world wide web that provides a way to integrate content. It allows machines to “understand” the content of sites by structuring it in such a way machines can interpret semantically structured knowledge.

Issues with Using an Outdated Web Approach
Currently, the internet is structured for humans to read and leaves no room for computers to interpret the content in a meaningful way. The current methods limit how information can be manipulated to benefit users and it gives content creators too much control over what reaches the eyes of those who use the internet.

Information retrieval has been a challenging task for search engines, whose goal is to provide quality results to users. Traditionally, search engines exploit the hyperlinks involved with a webpage and using page rank algorithms and a site’s HTML, search engines use information retrieval techniques to provide the most relevant information to the user and ranking this information by its relevancy to the user.

With hypertext markup language (HTML) there is no way to specify what items on a page are associated with a concept or entity. HTML does not allow for the representation of types within a concept, what and how information is tied together involved in a concept or entity, or how these conceptual items are distinct from other items.

A major issue of hyperlinks is that they can be linked to anything. Search engines assume content creators use links in a meaningful way, but that assumption gives users results that vary widely in relevancy, quality, and conceptual depth. These links do not reveal what the content means, only how the curator decides to connect content.

The world wide web needs to head in a new direction that goes beyond focusing on how content creators use HTML and hyperlinks and move towards a semantic web. Here, those that create or manage content can add meaning to their information documents that machines can interpret. When meaning can be translated across human and machine, it encourages the cooperation between the two. Consequently, search engine results will be more meaningful.

Semantics in the Semantic Web
A semantic web would make the internet more useful. Currently, search engine crawlers or crawling agents only index the information on web pages. Machines should be aware of the content within a page. For machines to know what to do when they encounter information, they must be able to interpret the meaning of the content or its semantics. With links between data and the embedding of semantic metadata, machines can understand particular concepts and the properties associated with them and what is shared across various concepts.

In the development of a semantic web, a knowledge representation must be encoded into a machine in order for semantics to be constructed. With structured data collections and computers reasoning capabilities, information can be defined and linked so that it allows computers to reason with, integrate and utilize this information across a multitude of technologies.

Current Challenges in Building Knowledge Representations
There are critical issues that arise in creating a global knowledge representation to be used by the semantic web. Specifically, knowledge representations require strict, well-defined definitions of the concept or entity that is being represented. Everyone (content curators and applications) need to share the same definition for commonly shared concepts, such as a car or a chair.

This is an issue because meaning is fuzzy. There will be overlap between distinct items within a concept, but they also will have their discrete properties that separate the two. Knowledge representation approaches need to account for the heterogeneity and dynamics involved in concepts, especially when dealing with such a vast amount of real-world data