Knowledge Update in Collaborative Knowledge Sharing Systems

Abstract

In a collaborative knowledge sharing system each source is associated with knowledge base system that participates in knowledge sharing with other sources. Acquaintances of sources build a collaborative knowledge sharing system or network in which each source is acquainted with other sources. The network of sources can be either acyclic or cyclic, meaning that they can contain acquaintance chains that are acyclic or cyclic. Updating knowledge in the sources involved in an acyclic logical network of sources is done by propagating an update from the originating source until the update reaches the leaves of the network. However, cyclic cases may create complexities due to conflicts that may arise between different versions of propagated updates. The author examines update propagation in both cyclic and acyclic networks. Moreover, the authors considers cases where a source is temporarily unavailable or offline. Here the author's propagation mechanism keeps track of every source even if the source is not available for a certain period of time until that source becomes available. Once a source comes back online the system must propagate the update destined to the returning sources to keep its knowledge consistent with other sources. The author has implemented this mechanism and evaluated it on a small collaborative knowledge sharing system.