December 29, 2006

Rich Tags Project

Eric told me about the Rich Tags project at mSpace. Rich Tags aims to improve the meaningfulness of tagging through lightweight visual and conceptual associations of a user's tags with "artifacts" in the application.

The act of tagging via association itself creates the meaning of the tag in both human and machine tractable terms. In effect, the act of semantic tagging effectively builds light weight ontologies.

Source: Rich Tags Project Description

December 9, 2006

David Huynh's Exhibit

Exhibit is the latest Project Simile "super data demo" originally developed by David Huynh. In the same vein as his Timeline, Exhibit is a simple client-side JSON data-interaction application that does not require any server-side code. Items are filtered and sorted into a main panel along the lines of a faceted browsing app.

Programmatically, an Exhibit data model can be queried much like an RDF graph: a query triple pattern is provided with either the subject or the object missing and values fitting that component in the triple pattern are returned. The query interface also allows for a set of values in either the subject or the object component of the query triple pattern; this proves useful for implementing faceted browsing. Source: Understanding Exhibit Expressions

Update: I just noticed Danny's post with a link to the project paper (which is definitely worth reading).

December 8, 2006

Exploring Social Annotations for the Semantic Web

Slashdot indirectly pointed me to a very useful paper that was presented last May at the 15th WWW Conference. In Exploring Social Annotations for the Semantic Web, Xian Wu, Lei Zhang, and Yong Yu describe an approach to improving the disambiguation of tags within social bookmarking service. The "co-occurrences" of tags logged by users with similar interests for a given url are analyzed to form knowledge categories and to derive the semantics of the free-form assertions made by users.

Representation of Semantics:

• We represent semantics of an entity (a web resource, a tag or a user) as a multi-dimensional vector where each dimension represents a special category of knowledge.
• Every entity can be mapped to a multi-dimensional vector, whose component measures the relativity between the entity and the category of knowledge. If one entity relates to a special category of knowledge, the corresponding dimension of its vector has a high score.
• The total knowledge of users, tags and resources are the same, we can represent them in the same multi-dimensional space, which we call conceptual space.

Source: Slides from the WWW2006 Presentation

As more objects in web apps become taggable by users, this approach can be used to determine which object properties users are interested in annotating, searching, and viewing. The UIs for lists and views could even be customized per user to emphasize the aggregated data of other users with similar interests.