Artificial intelligence: a new synthesis
Artificial intelligence: a new synthesis
The mathematics of statistical machine translation: parameter estimation
Computational Linguistics - Special issue on using large corpora: II
ACL '96 Proceedings of the 34th annual meeting on Association for Computational Linguistics
Statistical phrase-based translation
NAACL '03 Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology - Volume 1
Hi-index | 0.00 |
The paper describes basic properties of a sentence generator which requires minimal input information. Input is a set of unstructured semantic concepts, and the generator produces sentences which are compatible with this set by utilizing information from a statistical language model. Output is filtered by a simple context-free grammar. The system is trained on text from electronic medical records, and it is able to produce well-formed sentences in cases involving simple medication prescriptions and symptom descriptions. Basic complexity aspects of the problem are described, and suggestions for efficient implemented generators which manage to produce sentences within acceptable time limits, despite the complexity of the approach, are presented in the final sections.