Paper: Trainable Methods For Surface Natural Language Generation

ACL ID A00-2026
Title Trainable Methods For Surface Natural Language Generation
Venue Annual Conference of the North American Chapter of the Association for Computational Linguistics
Session Main Conference
Year 2000

We present three systems for surface natural lan- guage generation that are trainable from annotated corpora. The first two systems, called NLG1 and NLG2, require a corpus marked only with domain- specific semantic attributes, while the last system, called NLG3, requires a corpus marked with both semantic attributes and syntactic dependency infor- mation. All systems attempt to produce a grammat- ical natural language phrase from a domain-specific semantic representation. NLG1 serves a baseline system and uses phrase frequencies to generate a whole phrase in one step, while NLG2 and NLG3 use maximum entropy probability models to indi- vidually generate each word in the phrase. The sys- tems NLG2 and NLG3 learn to determine both the word choice and the word order of the phrase. We present e...