Error Analysis and the Role of Morphology

Marcel Bollmann, Anders Søgaard


Abstract
We evaluate two common conjectures in error analysis of NLP models: (i) Morphology is predictive of errors; and (ii) the importance of morphology increases with the morphological complexity of a language. We show across four different tasks and up to 57 languages that of these conjectures, somewhat surprisingly, only (i) is true. Using morphological features does improve error prediction across tasks; however, this effect is less pronounced with morphologically complex languages. We speculate this is because morphology is more discriminative in morphologically simple languages. Across all four tasks, case and gender are the morphological features most predictive of error.
Anthology ID:
2021.eacl-main.162
Volume:
Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume
Month:
April
Year:
2021
Address:
Online
Editors:
Paola Merlo, Jorg Tiedemann, Reut Tsarfaty
Venue:
EACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1887–1900
Language:
URL:
https://aclanthology.org/2021.eacl-main.162
DOI:
10.18653/v1/2021.eacl-main.162
Award:
 Best Long Paper
Bibkey:
Cite (ACL):
Marcel Bollmann and Anders Søgaard. 2021. Error Analysis and the Role of Morphology. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1887–1900, Online. Association for Computational Linguistics.
Cite (Informal):
Error Analysis and the Role of Morphology (Bollmann & Søgaard, EACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.eacl-main.162.pdf
Code
 coastalcph/eacl2021-morpherror