The concept of a Semantic Web offers promising possibilities for the future development of the Internet. Through the use of ontology languages and logic engines, the Internet can incorporate a layer of intelligent information available to all websites to tap into. Web applications will be able to make intelligent decisions through the use of such information. Interesting examples include human-like problem solving through smarter search and deductions; and online psychological research through semantic information available from social networks.
However, the process to establish a completely semantic layer of the web is fraught with hurdles. The most prominent of these hurdles is the need for machine-readable representations of data. Although frameworks such as RDF and OWL have been adopted, one still requires human input to convert mundane pieces of information into these formats. With the rapidly expanding scale of content on the Internet at present, it would be a near-impossible task to convert all the content into the standardized RDF-based formats.
Further, basic human intelligence that is termed “common sense” is difficult to emulate on a machine. Research in Artificial Intelligence (AI) continues to develop a satisfactory model that is at par with it. This sort of intelligence is difficult to classify into ontologies and axioms, let alone represent in any of the ontology languages.
For semantic web to develop into an embedded part of the World Wide Web, computers require AI capabilities to convert content found on the Internet into semantic data. If this process is successfully automated, it can be automatically applied to any new content generated on the Internet, thus ensuring that all information is available to machines as semantically logical and understandable data.
IBM’s Watson is a recent example of a machine that utilizes a database of “triples” for search queries. More important for Watson, however, is the ability to properly understand a question before beginning to search it. As explained in detail in this paper published in the AI Magazine, Association for the Advancement of Artificial Intelligence, September 2010 (http://aaaipress.org/ojs/index.php/aimagazine/article/download/2303/2165), Watson converts the question into semantic objects in order to extract relationships between them. The ability to create a semantic view on its own is the key to its success.
Online social networks provide an ideal environment for collecting and automatically developing semantic information. Using the APIs usually provided by social networking websites, computers can construct OWL ontologies fairly easily. Example, the Friends of a Friend project is currently supported by several social networking websites. Thus, tapping into social networks is an excellent starting point for automatically generating semantic information out of published content on the web. Moreover, for the limited scope of describing and documenting connections between people, the use of semantic web over online social networking is sufficient.
In conclusion, I believe the concept of semantic web is promising, but requires the process of generating semantic information to be automatized, and coupled with comparative artificial representations of humanoid intelligence. Going ahead, the advancements in AI will definitely promote this field. Yet, the more humble goal of building a global social network may be practical and quite achievable.