Exposing Latent Information in Folksonomies for Reasoning

A new family of what is known as Web 2.0 applications is currently emerging on the Web as means to create and share web resources. These applications include user-centric publishing and knowledge management platforms like Wikis and Blogs. Social resource sharing systems have acquired several million users in just few years. As the use of folksonomy proliferate users feel the need for more structure for better organizing their resources and enhance resource discovery and retrieval. We believe that exposing such latent structure will enhance user experience in a wide range of applications including situation awareness. We conjecture that a complete solution that is able to deduce latent structures in user vocabulary must account for controlled and uncontrolled vocabularies as well as the three elements of the social tagging triangle: people, tags, and resources. Controlled vocabularies could be as limited as domain ontology but it could also be as wide as Wordnet, Propbank or Framenet. In our view, a successful system design an implementation should not restrict to a specific vocabularies. Instead we propose to develop general solutions that extract latent structures in folksonomies through mapping to semantically richer controlled vocabularies.