SIGMORPHONUniMorph 2022 Shared Task 0: Modeling Inflection in Language Acquisition

Jordan Kodner, Salam Khalifa


Abstract
This year’s iteration of the SIGMORPHONUniMorph shared task on “human-like” morphological inflection generation focuses on generalization and errors in language acquisition. Systems are trained on data sets extracted from corpora of child-directed speech in order to simulate a natural learning setting, and their predictions are evaluated against what is known about children’s developmental trajectories for three well-studied patterns: English past tense, German noun plurals, and Arabic noun plurals. Three submitted neural systems were evaluated together with two baselines. Performance was generally good, and all systems were prone to human-like over-regularization. However, all systems were also prone to non-human-like over-irregularization and nonsense productions to varying degrees. We situate this behavior in a discussion of the Past Tense Debate.
Anthology ID:
2022.sigmorphon-1.18
Volume:
Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology
Month:
July
Year:
2022
Address:
Seattle, Washington
Venue:
SIGMORPHON
SIG:
SIGMORPHON
Publisher:
Association for Computational Linguistics
Note:
Pages:
157–175
Language:
URL:
https://aclanthology.org/2022.sigmorphon-1.18
DOI:
10.18653/v1/2022.sigmorphon-1.18
Bibkey:
Cite (ACL):
Jordan Kodner and Salam Khalifa. 2022. SIGMORPHON–UniMorph 2022 Shared Task 0: Modeling Inflection in Language Acquisition. In Proceedings of the 19th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 157–175, Seattle, Washington. Association for Computational Linguistics.
Cite (Informal):
SIGMORPHON–UniMorph 2022 Shared Task 0: Modeling Inflection in Language Acquisition (Kodner & Khalifa, SIGMORPHON 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.sigmorphon-1.18.pdf
Code
 sigmorphon/2022inflectionst
Data
CELEX