Towards parallel nonmonotonic reasoning with billions of facts

Abstract

We are recently witnessing an explosion of available
\ndata from the Web, government authorities, scientific
\ndatabases, sensors and more. Such datasets could benefit
\nfrom the introduction of rule sets encoding commonly
\naccepted rules or facts, application- or domainspecific
\nrules, commonsense knowledge etc. This raises
\nthe question of whether, how, and to what extent knowledge
\nrepresentation methods are capable of handling the
\nvast amounts of data for these applications. In this paper,
\nwe consider non-monotonic reasoning, which has
\ntraditionally focused on rich knowledge structures. In
\nparticular, we consider defeasible logic, and analyze
\nhow parallelization, using the MapReduce framework,
\ncan be used to reason with defeasible rules over huge
\ndata sets. Our experimental results demonstrate that defeasible
\nreasoning with billions of data is performant,
\nand has the potential to scale to trillions of facts.