Facilitating Lightweight Consensuses in P2P Environments

Download

PDF - Accepted Version Restricted to Registered users only Download (161Kb)

Description/Abstract

Traditional ontology mapping techniques are not strictly applicable in a dynamic and distributed environment (e.g. P2P and pervasive computing) in which on-the-fly alignments are sought after. We propose a theoretical framework that collaborates the logic formalisms with collaboratively created web repositories. A graph clustering algorithm is to discover, from concept definitions, the “feature” vectors that uniquely identify concepts; web repositories are used to understand the implications of these features. Such a combination solidifies an on-demand and approximate mechanism that emerges a context-dependent and task specific consensus among heterogeneous participants of an information exchange task.