Discovery and Reasoning

The use of discovery and reasoning makes this web services platform semantically enabled. We use Pellet, from the Mindswap Group for our reasoning engine.

Reasoning:
Resource description graphs registered ( see this url ) to the discovery server are first validated for OWL DL consistency. Then statements from the resource description graph along with any ontological terms the graph imports are reasoned over by Pellet. All explicit statements from the resource description graph, explicit statements from imported term definitions and implicit statements from the reasoner are then added to the knowledge base and persisted to a database. We used a set of resources description graphs, term definitions and query graphs from the LIS ( Legume Information System )team to validate the requirements for the reasoning functionality. The package of use case graphs can be downloaded here. The query graphs are submitted to the discovery server via our graph search tool here . Additionally, this document describes the reasoning design in more detail.

Discovery:
We are currently developing a human interactive interface to present queries to the discovery server that exercises the underlying semantics and reasoning. Initially, we will perform a combination of token matching and ontology term matching to find resources. We will leverage the underlying semantics within the knowledge base by prioritizing the set of results returned for a given search. This document provides more detail to the design of the search functionality.