Agents

An agent is essentially a piece of software that collects web content from diverse sources, processes the information and can exchange results with other agents. Since the semantic web uses the power of semantics this means that even agents that were not designed to work together can still transfer data among themselves.

One important aspect of  an agents functioning will be the exchange of "proofs". Written in the Semantics Webs unifying language "the language that expresses logical inferences made using rules and information such as those specified by ontologies" Scientific America.

This means that if you are not sure about the answer to a query provided by an agent, then it can verify its data and translate the proof from the agents internal reasoning into the Webs unifying language.  An inference engine within your computer readily verifies the information.

A simple example of Agents exchanging proofs.

Agent to agent communication 

Of course Agents will have to communicate with each other all the time, and if they are all using different ontologies, which in turn can all be linked to other ontologies, this can get complicated. The figure above shows what a small section of the web might look like.  The small boxes represent agents and the larger boxes represent ontologies. In theory any two Agents should be able to communicate with each other by finding a common ontology and mapping into it. This would only work however if the web were perfect unfortunately it is more likely that the ontologies will be imperfectly or only partially mapped.  

Agents Will make the Semantic Web extremely flexible Since the consumer and producer agents will be able to share understandings through exchanging ontologies. Agents will even be able to gain new capabilities by absorbing new ontologies. A typical process on the Semantic Web might involve the setting up of a value chain whereby information is passed from one agent to another each one "adding value" to eventually make the final product as requested by the user. This will mean however that some agents will need to exploit Artificial intelligence to be able to do this on demand.  

Previous chapter                                                                         Next chapter