This week, the Knowledge Engine team reached a milestone in the development of the distributed reasoner.

The distributed reasoner is capable of automatically combining knowledge that is spread over different systems. The team has adapted Apache Jena's rule system's test suite to guarantee a baseline of functionality.

After designing a special algorithm for the reasoner and rebuilding all the functionality such as automatic knowledge gap discovery and explainability, the distributed reasoner is now more capable than ever.

reasoning graph consisting of related rules

A reasoning graph consisting of related rules.

The algorithm uses multiple passes over the rules in the reasoning graph, and first propagates filters through the graph's rules so that only the required data is considered. This is followed by result propagation where the derived facts are spread among the rules in the graph.

With this improved reasoner, we have a foundation for many exciting new possibilities, offering even more flexibility for knowledge bases that use the Knowledge Engine.