Towards Swarm-based Federated Web Knowledgebases

Abstract

Internet knowledgebases are more and more described using ontological vocabularies. However, efficient coherent solutions for both federated storage and reasoning upon widely distributed web repositories are rarely explored. We present a self-organized, distributed storage and reasoning approach based on swarm intelligence exploiting the strong applicability of swarm algorithms to distributed environments. While centralized reasoning has problems in scalability, swarm intelligence has the potential to reach Web scalability. Our concept comprises two layers of swarm algorithms - storage and reasoning layer - for information storage and inference of new statements, respectively. We present for the former our concepts with first evaluation results and give for the latter a general swarm-based concept for forward-chaining assertional reasoning in the description logic ALC.