Semantic-based knowledge systems
From data to knowledge: the semantic web
New services developed by European researchers could help turn the World Wide Web into a vast relational database allowing surfers to extract relevant data for any topic imaginable.
The researchers are in the process of helping develop what is called the ‘semantic web’, which promises to give users access to actionable or useable knowledge instead of just vast quantities of information.
The internet is the greatest store of information assembled in human history. But its very size can prevent it from becoming a truly useful tool. There is so much information that users find it difficult to filter out the inaccurate and irrelevant.
We cannot see the wood for the trees. We have lots of data, but we must exclude most of it. Call it the knowledge gap, the distance between what we can access and what we need.
That gap is closing. Researchers at the Semantic Knowledge Technologies (SEKT) project have developed a robust and comprehensive system to identify the content, relevance and quality of information contained on your computer, your local network or the World Wide Web.
The age of knowledge
Essentially, the SEKT system can identify the meaning and significance of data and the semantic content, and can thus help users extract quality information rather than just quantity.
For example, take John. He is a stockbroker with 200 clients. He is familiar with six of them. With the SEKT suite working in the background, John's information technology (IT) systems can scan his incoming email to highlight relevant text and phrases, supporting him in servicing his clients.
If clients want to shift their strategy from capital growth to income earning, John's IT system can start presenting useful options to him, so he can best advise his clients. Or the system could scan incoming newsfeeds, highlighting stories that impact his clients’ stock portfolio.
The computer is analysing the data faster than John could sort it, giving him access to vast resources.
The example is simple, and it is difficult to guess which services will be available when the technologies become widespread. But the potential is limitless.
How it works
Any computer can identify, with a fair probability of success, the content of a file by name. But computers are not equipped to understand what the content means, the context or the data’s relevance to other pieces of data.
Semantic technologies fill this gap. They typically define what a piece of data actually means. They are based on ontologies, or dictionaries, of metadata.
Metadata is simply information about a piece of information. Metadata describes content in a way that is machine-readable, allowing computers to identify the context of data. Machines can then search for relevant and timely knowledge.
The metadata defines what a piece of information means. The ontologies list the definitions and typically they are organised by domain, subjects like health, business organisation, or government and so on.
When you combine ontologies and metadata with computers and networking, you get the semantic web, a machine-readable internet.
SEKT’s researchers have also developed a way to generate ontologies semi-automatically, and to manage ontologies that already exist. They have produced tools to use these ontologies in standard IT systems, and created an open source application to tie everything together.
To validate their work, SEKT’s researchers tested the system in such areas as law, digital library management and IT consultancy. These are all areas that deal with vast quantities of qualitative information. The tests were done in Spanish, German and English, respectively.
Fields of science
Call for proposal
See other projects for this call
Funding SchemeIP - Integrated Project