Survey: Evaluating the Quality of Texts Produced by NLP Systems

Authors

  • Ehud Reiter Department of Computing Science, University of Aberdeen

Abstract

I survey techniques and experimental designs used to evaluate the quality of texts produced by NLP systems, including machine translation, natural language generation, and summarisation.  I present evaluation as a type of scientific hypothesis testing, and include in this survey papers from the broader scientific community as well as papers from the NLP community.

Author Biography

  • Ehud Reiter, Department of Computing Science, University of Aberdeen
    Professor of Computing Science at the University of Aberdeen, and Chief Scientist of Arria NLG

Published

2024-12-05

Issue

Past Publications

Section

Survey Article