Subject
Knowledge Representation and the Semantic Web
Describing a domain so that a machine can draw conclusions from the description. Subsumption between concepts is a partial order, so a hierarchy is not a tree and need not have a meet or a join for every pair, which is why description logics supply constructors rather than relying on named concepts. A classifier then returns everything the axioms entail, often more than was drawn, and its silence means the ontology entails nothing either way rather than that the answer is no.
Learning paths
Knowledge Representation and the Semantic Web
Reading a concept hierarchy as an order, and predicting what a reasoner concludes from a set of axioms.
What this subject develops
Predict what a knowledge model entails
Treat a concept hierarchy as an ordered structure, compute its bounds, and predict which subsumptions a classifier derives and which it leaves underived.
See detailed outcomes
Predict what a knowledge model entails
- The learner can verify that subsumption is a partial order, locate meets and joins where they exist and identify pairs that have none, predict the subsumptions a classifier derives from asserted axioms, and trace an inference failure to a missing axiom or to the open-world reading rather than to the tool.
Browse Knowledge Representation and the Semantic Web reference · Practice Knowledge Representation and the Semantic Web