The automatic initialization of an object-oriented knowledge base

ABSTRACT

This paper describes mechanisms for automatically building an initial set of knowledge base objects by processing natural language text in a restricted domain. A separate component assigns syntactic and domain-specific semantic tags to each of the words in the text. The syntactic and semantic tags are used to generate names of objects to be added to specific semantic classes in the knowledge base. An algorithm to ensure that the knowledge base objects created are not redundant is specified. The definition and organization of metaknowledge structures is also discussed. To ensure that the name generation and knowledge management techniques are domain-independent, any domain-specific information is isolated in a knowledge base definition file.

REFERENCES

• 1: Jim Barnett , Kevin Knight , Inderjeet Mani , Elaine Rich, Knowledge and natural language processing, Communications of the ACM, v.33 n.8, p.50-71, Aug. 1990 [doi>10.1145/79173.79177]
• 2: Boggess, L., Agarwal, R., and Davis, R. 1991. "Disambiguation of prepositional phrases in automatically labelled technical text." In Proceedings of AAAI 91, Vol. 1 : 155-59.
• 3: Julia E. Hodges , Jose L. Cordova, Automatically Building a Knowledge Base Through Natural Language Text Analysis, Mississippi State University, Mississippi State, MS, 1991
• 4: Padgham, L. 1988. "A model and representation for type information and its use in reasoning with defaults." In Proceedings of AAAI 88 : 409-14.