We Need to Talk about Standard Splits

Kyle Gorman, Steven Bedrick


Abstract
It is standard practice in speech & language technology to rank systems according to their performance on a test set held out for evaluation. However, few researchers apply statistical tests to determine whether differences in performance are likely to arise by chance, and few examine the stability of system ranking across multiple training-testing splits. We conduct replication and reproduction experiments with nine part-of-speech taggers published between 2000 and 2018, each of which claimed state-of-the-art performance on a widely-used “standard split”. While we replicate results on the standard split, we fail to reliably reproduce some rankings when we repeat this analysis with randomly generated training-testing splits. We argue that randomly generated splits should be used in system evaluation.
Anthology ID:
P19-1267
Volume:
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics
Month:
July
Year:
2019
Address:
Florence, Italy
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2786–2791
URL:
https://www.aclweb.org/anthology/P19-1267
DOI:
10.18653/v1/P19-1267
Bib Export formats:
BibTeX MODS XML EndNote
PDF:
https://www.aclweb.org/anthology/P19-1267.pdf