A model of natural language generation based on an underlying language-free representation of meaning is described. A program based on this model is able to produce sentence paraphrases which demonstrate understanding with respect to a given context. This generator operates in conjunction with a natural language analyzer and a combined memory and inference model. In generating sentences from meaning structures, the program employs both the information retrieval and deduction capabilities of the memory model. The model encompasses several diverse classes of linguistic knowledge, which include: (1) executable tests of conceptual properties stored in discrimination nets; (2) information relating conceptual to syntactic roles, stored in a word-sense dictionary, and (3) surface grammatical knowledge, stored in a formal grammar.
Sentence paraphrasing from a conceptual base
The Latest from CACM
Shape the Future of Computing
ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.
Get InvolvedCommunications of the ACM (CACM) is now a fully Open Access publication.
By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.
Learn More
Join the Discussion (0)
Become a Member or Sign In to Post a Comment