Knowledge Graphs Versus Conceptual Graphs

摘要

Two knowledge representation formalisms based on graphs are compared particularlywith respect to their use for natural language semantics. A natural language model in terms of knowledge graphs, that has three main processes: activation, contraction, and generalization is described. The features of conceptual graphs: the type hierarchy, the conformity relation, the canonical derivation rules, the generalization relation, and the important maximal join are all proved to be present in the knowledge graph model.