This project topic focuses on technologies and applications at the intersection of the Semantic Web, Web 2.0 and Grid/Web services.
Projects will be closely integrated with the work of the PolicyGrid research project in Computing Science.
I am interested in how Web based software systems can be made easier to use and "smarter" by integrating aspects of AI and Web technology; much of my current work is related to the so-called Social Semantic Web. This is envisaged as a new generation of applications that combine the the data flexibility and portability that is characteristic of the Semantic Web, and the scalability and authorship advantages of Web 2.0 applications.
The Semantic Web encapsulates the vision of a Web enriched by annotations which allow both software and humans to use the content by making the Web to some degree "machine-understandable". Core technologies include the Resource Description Framework (RDF) - a foundation for processing XML-based metadata which provides interoperability between applications that exchange machine-understandable information on the Web; the Web Ontology Language (OWL) - used to explicitly represent the meaning of terms in vocabularies and the relationships between those terms (known as an ontology). Several Semantic Web applications already exist, including: a Semantic Web version of the Firefox browser (called PiggyBank) and an evening scheduler for a night out in Aberdeen - GraniteNights.
The term Web 2.0 is now routinely used to refer to a group of Internet-based services — such as social networking sites, wikis, communication tools, and folksonomies — that let people collaborate and share information online in ways previously unavailable. Central to the Web 2.0 vision is the idea of "harnessing the collective intelligence" of a community of users to offer new kinds of Web content and services. Sites such as Wikipedia, Flickr and MySpace are all examples of Web 2.0 services. Folsonomies are an interesting idea (used by many Web 2.0 sites); simply put, a folksonomy is a set of community generated, open-ended labels that categorize content such as Web pages, photographs, or other Web accessible resources. A folksonomy is different to an ontology (see above) in that the authors of the labeling system are often the main users (and sometimes originators) of the content to which the labels (known as "tags") are applied.
The PolicyGrid project has developed the ourSpaces Web environment as a Social Semantic Web demonstrator. ourSpaces (see screenshot below) integrates common features seen in social networking sites, with rich metadata support (based upon integration of ontologies and folksonomies) and mechanisms for sharing digital artefacts. ourSpaces has been designed around a component based architecture, so that developers can produce new components for seamless integration into the interface. Our ultimate aim is to explore how we can enhance all seven of the building blocks of a Web2.0 application - identity, relationships, groups, sharing, presence, conversations, and reputation - using semantics. For further details see our paper - Building a Social Semantic Web for eScience.
All of these projects relate to our work on integrating aspects of social and semantic technology, many in the context of ourSpaces. If you are interested in working in this area, you are encouraged to chat to Richard Reid (a research assistant on the PolicyGrid project, and one of last year's final year students) - as he has developed much of the ourSpaces infrastructure.
There are several potential sub-projects here and suggestions include:
Exploring how presence can be represented and used within a Social Semantic Web environment. While presence can be thought of as a representation of communication states such as "free for chat", "busy", "away", "do not disturb", "out to lunch - there are also many other possible interpretations, e.g. based on a user's location, social contacts, etc.
Development of novel interfaces to allow browsing/visualisation of scientific metadata presented via ontologies and folksonomies, e.g. via Google maps, timelines, etc.
Exploring how analysis of folksonomies (created by a user community) can be used to refine/extend existing OWL ontologies. For example, can frequently occurring tags be considered as potential new classes within an OWL ontology?
Developing text-mining tools which can analyse text documents to suggest candidate folksonomy tags and/or ontology classes to be used as metadata. Such a facility could be integrated into the existing ourSpaces environment.
These are just a selection - there are many others (all of which involve working with members of the PolicyGrid project team).
If you would like to make an appointment to discuss project ideas with Dr Edwards, please send an email to:
p.edwards@abdn.ac.uk indicating your availability, and he'll get back to you with an appointment slot. Dr Edwards is Head of the Graduate School for the College of Physical Sciences and he does not see students without an appointment.
You should not choose this topic area without talking to Dr Edwards first!
A few papers that will give an insight into some of the issues:
Folksonomies - Tidying up Tags?
DLib Magazine, January 2006
M Guy & E Tonkin
[html]
Building a Social Semantic Web for eScience
Submitted to AAAI Spring Symposium on Social Semantic Web: Where Web 2.0 Meets Web 3.0
P Edwards, R Reid, A Chorley, F Hielkema, R Mitra & E Pignotti
[pdf]
GraniteNights - A Multi-Agent Visit Scheduler Utilising Semantic Web
Technology
Proceedings of Seventh International Workshop on Cooperative Information Agents (CIA 2003)
G A Grimnes, S Chalmers, P Edwards & A D Preece
[pdf]