Natural Language Generation for Text-to-Text Applications Using an Information-Slim Representation

Radu Soricut

I propose a representation formalism and algorithms to be used in a new language generation mechanism for text-to-text applications. The generation process is driven by both text-specific information encoded via probability distributions over words and phrases derived from the input text, and general language knowledge captured by n-gram and syntactic laguage models.

Subjects: 13. Natural Language Processing

Submitted: Apr 4, 2005


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.