Acquiring Correct Knowledge for Natural Language Generation

Research output: Contribution to journalArticle

44 Citations (Scopus)

Abstract

Natural language generation (NLG) systems are computer software systems that produce texts in English and other human languages, often from non-linguistic input data. NLG systems, like most AI systems, need substantial amounts of knowledge. However, our experience in two NLG projects suggests that it is difficult to acquire correct knowledge for NLG systems; indeed, every knowledge acquisition (KA) technique we tried had significant problems. In general terms, these problems were due to the complexity, novelty, and poorly understood nature of the tasks our systems attempted, and were worsened by the fact that people write so differently. This meant in particular that corpus-based KA approaches suffered because it was impossible to assemble a sizable corpus of high-quality consistent manually written texts in our domains; and structured expert-oriented KA techniques suffered because experts disagreed and because we could not get enough information about special and unusual cases to build robust systems. We believe that such problems are likely to affect many other NLG systems as well. In the long term, we hope that new KA techniques may emerge to help NLG system builders. In the shorter term, we believe that understanding how individual KA techniques can fail, and using a mixture of different KA techniques with different strengths and weaknesses, can help developers acquire NLG knowledge that is mostly correct.

Original languageEnglish
Pages (from-to)491-516
Number of pages25
JournalJournal of Artificial Intelligence Research
Volume18
Publication statusPublished - Jan 2003

Keywords

  • SMOKING-CESSATION LETTERS

Cite this

Acquiring Correct Knowledge for Natural Language Generation. / Reiter, Ehud Baruch; Sripada, Gowri Somayajulu; Robertson, R.

In: Journal of Artificial Intelligence Research, Vol. 18, 01.2003, p. 491-516.

Research output: Contribution to journalArticle

@article{ac431a799c464816902197930abb2f0c,
title = "Acquiring Correct Knowledge for Natural Language Generation",
abstract = "Natural language generation (NLG) systems are computer software systems that produce texts in English and other human languages, often from non-linguistic input data. NLG systems, like most AI systems, need substantial amounts of knowledge. However, our experience in two NLG projects suggests that it is difficult to acquire correct knowledge for NLG systems; indeed, every knowledge acquisition (KA) technique we tried had significant problems. In general terms, these problems were due to the complexity, novelty, and poorly understood nature of the tasks our systems attempted, and were worsened by the fact that people write so differently. This meant in particular that corpus-based KA approaches suffered because it was impossible to assemble a sizable corpus of high-quality consistent manually written texts in our domains; and structured expert-oriented KA techniques suffered because experts disagreed and because we could not get enough information about special and unusual cases to build robust systems. We believe that such problems are likely to affect many other NLG systems as well. In the long term, we hope that new KA techniques may emerge to help NLG system builders. In the shorter term, we believe that understanding how individual KA techniques can fail, and using a mixture of different KA techniques with different strengths and weaknesses, can help developers acquire NLG knowledge that is mostly correct.",
keywords = "SMOKING-CESSATION LETTERS",
author = "Reiter, {Ehud Baruch} and Sripada, {Gowri Somayajulu} and R. Robertson",
year = "2003",
month = "1",
language = "English",
volume = "18",
pages = "491--516",
journal = "Journal of Artificial Intelligence Research",
issn = "1076-9757",
publisher = "Morgan Kaufmann Publishers, Inc.",

}

TY - JOUR

T1 - Acquiring Correct Knowledge for Natural Language Generation

AU - Reiter, Ehud Baruch

AU - Sripada, Gowri Somayajulu

AU - Robertson, R.

PY - 2003/1

Y1 - 2003/1

N2 - Natural language generation (NLG) systems are computer software systems that produce texts in English and other human languages, often from non-linguistic input data. NLG systems, like most AI systems, need substantial amounts of knowledge. However, our experience in two NLG projects suggests that it is difficult to acquire correct knowledge for NLG systems; indeed, every knowledge acquisition (KA) technique we tried had significant problems. In general terms, these problems were due to the complexity, novelty, and poorly understood nature of the tasks our systems attempted, and were worsened by the fact that people write so differently. This meant in particular that corpus-based KA approaches suffered because it was impossible to assemble a sizable corpus of high-quality consistent manually written texts in our domains; and structured expert-oriented KA techniques suffered because experts disagreed and because we could not get enough information about special and unusual cases to build robust systems. We believe that such problems are likely to affect many other NLG systems as well. In the long term, we hope that new KA techniques may emerge to help NLG system builders. In the shorter term, we believe that understanding how individual KA techniques can fail, and using a mixture of different KA techniques with different strengths and weaknesses, can help developers acquire NLG knowledge that is mostly correct.

AB - Natural language generation (NLG) systems are computer software systems that produce texts in English and other human languages, often from non-linguistic input data. NLG systems, like most AI systems, need substantial amounts of knowledge. However, our experience in two NLG projects suggests that it is difficult to acquire correct knowledge for NLG systems; indeed, every knowledge acquisition (KA) technique we tried had significant problems. In general terms, these problems were due to the complexity, novelty, and poorly understood nature of the tasks our systems attempted, and were worsened by the fact that people write so differently. This meant in particular that corpus-based KA approaches suffered because it was impossible to assemble a sizable corpus of high-quality consistent manually written texts in our domains; and structured expert-oriented KA techniques suffered because experts disagreed and because we could not get enough information about special and unusual cases to build robust systems. We believe that such problems are likely to affect many other NLG systems as well. In the long term, we hope that new KA techniques may emerge to help NLG system builders. In the shorter term, we believe that understanding how individual KA techniques can fail, and using a mixture of different KA techniques with different strengths and weaknesses, can help developers acquire NLG knowledge that is mostly correct.

KW - SMOKING-CESSATION LETTERS

M3 - Article

VL - 18

SP - 491

EP - 516

JO - Journal of Artificial Intelligence Research

JF - Journal of Artificial Intelligence Research

SN - 1076-9757

ER -