Agreement is overrated: A plea for correlation to assess human evaluation reliability

Jacopo Amidei, Paul Piwek, Alistair Willis. Agreement is overrated: A plea for correlation to assess human evaluation reliability. In Kees van Deemter, Chenghua Lin, Hiroya Takamura, editors, Proceedings of the 12th International Conference on Natural Language Generation, INLG 2019, Tokyo, Japan, October 29 - November 1, 2019. pages 344-354, Association for Computational Linguistics, 2019. [doi]

Abstract

Abstract is missing.