Knowledge sharing using Semantic Web technologies

Author: Nick Kings

nick.kings@bt.com

 

1          Project Overview

SEKT (Semantic Enabled Knowledge Technologies) http://sekt.semanticweb.org aims to develop the essential semantic knowledge technologies to enable the “European Knowledge Society”. The vision of SEKT is to develop and exploit new knowledge technologies which include means for automated knowledge extraction, knowledge packaging and delivering according to user profiles as well as semantic-based knowledge analyses and matching for user-driven knowledge push. SEKT is an EU 6th Framework Integrated project.

 

A major barrier to a widespread use of knowledge management systems in industry and organizations arises from the necessary overhead of knowledge modelling and annotation.  SEKT will address these and other challenges by an interdisciplinary approach focussing on substantially reducing the overhead of knowledge modelling and annotation of sources.  This will be done by integrating Ontology & Metadata Technology (OMT), Human Language Technology (HLT), and Knowledge Discovery (KD) into a uniform and scalable framework that supports the integrated learning and management of ontologies and metadata in a (semi-) automatic way.

 

The three technologies will be combined to produce semi-automatic tools for the creation of ontologies, the population of those ontologies with meta-data, and the maintenance and evolution of the ontologies and associated metadata.  A knowledge access tool will also be developed to make use of the ontologies and associated metadata to meaningfully access the knowledge.  Middleware software, with published APIs, will enable the suite of tools to interoperate. Specifically, the use of ontologies and metadata underlies the SEKT components, and the whole approach; human language technology will be used to extract metadata; knowledge discovery will be used to semi-automatically learn and evolve ontologies.

 

2          Technical Background

Much of current research has focussed upon the individual and their use of knowledge. Profiles of users are built up model aspects of a user’s behaviour, and subsequently infer what information that user may want to see. From the research surveyed, a number of themes can be identified:

·        Dynamic context - roles

A knowledge sharing tool should be designed with an awareness of the tasks that a user is trying to achieve. In the simplest case, a user could visit an information portal, as part of their day-to-day tasks. As a user moves through an information portal, information associated with the current page can be presented. In this way, users can perform complex context-based searches without being aware that is being undertaken on their behalf. The context here is derived from the portal’s underlying ontology [1], [2], [3], [4], [5], [6].

Away from an information portal, the task of inferring a user’s context is far more difficult. A user may undertake a range of different tasks, each using a different set of tools to gather and collate information, such as email and a web browser. Haystack [7] seeks to overcome this problem by collating all information sources into one place, but that may not be feasible in all circumstances. Research is needed to determine a user’s current focus, and deliver that information in the most appropriate fashion.

·        Implicit sharing

An important question to answer is how to make the sharing of knowledge implicit in the normal work activities of a user? Activities such as Jasper [8], [9], OntoShare [10], [11] and Eureka [12], [13], [14] involve some effort to made on the part of the user to share information, but users share information that is already seen to be of value; in approaches such as SWAP [15], [16], [17], Social Web Cockpit [18], [19], [20], [21] and “Stuff I’ve Seen” [22], [23] derive as much information as possible from the user’s interactions with other software tools.

If information is shared explicitly, the software tool must be closely coupled with the day to day working of an individual, rather than an additional process. Thus, making knowledge sharing implicit is important, as it lowers the barriers for a user to make a contribution, however, if everything is being monitored, how does the software detect what information is significant or could be applicable in a new situation?

·        End user creation of material - governance

A significant amount of research is being undertaken into peer-to-peer knowledge sharing, as well as the gathering and collation of information from wide disparate sources. In a strictly hierarchical organisation, it is possible to find the definitive source of information. However, in a more distributed organisation, or a complex community, a user may seek advice from many different sources, but how is it possible to verify the governance or authority of a particular piece of advice?

·        Communities

Ontologies are being generated based upon an individual’s use of information, and how that information fits into the wider community, but little research is being undertaken on how to understand an individual fits within a community. Using Schniederman’s model [24] implies that the nature of the community has an impact upon the nature of the information shared. Careful consideration and research needs to be undertaken to understand how to share information between communities.

Knowledge sharing tools combine the functions of searching for and distributing information. As a user requires information to undertake a task, information relevant to that task can be located. Underpinning knowledge sharing tools is the premise that someone in the user’s wider community has already created relevant information (explicit knowledge transfer) or someone is able to provide help or advice (tacit knowledge transfer). Knowledge sharing tools use profiles of the users to identify and route information to people, as and when that information is required.

Ontologies have the potential to underpin and enable large scale knowledge sharing. Understanding a user’s context is the central research problem, which can be broken down into several areas:

·        The use of ontologies to support expertise location. In a large community, it is crucial that a user is able to validate another person’s reputation and skill in a particular subject.

·        The facilitation of distributed generation of meta-data. Each person is able to annotate, comment and manage their own information sources; this meta-data should be reconciled with the community’s wider information sources.

·        The identification of communities of interest, within wider communities. It is feasible to identify sets of people with common sets of interests, but it is crucial to understand how these groups form and can be maintained.

·        The understanding of changing user roles. The profile should be able to represent the fact that a person undertakes many different roles and responsibilities; the profile should represent both long term and short term interests.  Knowledge sharing tools must deliver information in a way that is relevant to both the user and to the task in hand. 

·        The underlying ontology will be used to identify implicit user needs, and be able to fetch information in advance of a user’s explicit query.

·        Increasing quantities of information is being stored in RSS and FOAF formats. Knowledge sharing tools should be able to make use of and exploit the semantic information contained therein, in order to validate items such as a user’s reputation and expertise.

SEKT will develop methods that monitor a user’s information usage, in order to profile his or her information needs. In turn, the tools will relate his or her information needs to the wider community’s ontology. Having derived a profile, the issue is to intelligently present relevant information, and route information between two, or more, members of the community. By classifying information against an ontology, facilities will be provided that augment each community member’s memory, and enhance recall of information at later points in time.

3          Research Position

As part of the SEKT project, Nick is undertaking a PhD, due to start in September, 2004; attendance at the FOAF Workshop is seen as an important first step in gaining an awareness of the current state of the art, in this technical domain.

The PhD will investigate the viability of using an ontology based peer-to-peer system. By providing the opportunity to closely monitor a user’s activity, it should be possible to identify changes of focus and possibly changes of role. To be a success, the peer-to-peer sharing application will derive a richer picture of a user’s context, based on currently important documents, as well as information derived from the activities from friends and colleagues. Research will focus upon the use of FOAF to provide a semantic overlay on a peer-to-peer network, in order to identify important participants and other information sources.

4          References