A Multiagent Text Generator with Simple Rhetorical Abilities
Authors
Abstract
The research presented in this paper is aimed at improving thestylistic quality of the texts generated by the PRINCE sys-
tem, a natural language generation application designed for
building texts for simple fairy tales. This is done by exploit-
ing the potential of a lexical resource - such as WordNet -
and structure mapping algorithms to enhance the output texts
with simple rhetorical tropes such as comparison, metaphor,
and analogy. The enhanced system follows the OAA multia-
gent system architecture, with various agents cooperating to
achieve the final texts. The results of the enhanced version are
presented and discussed. Finally, some problems and possi-
ble extensions are described.