When Irregularity Helps: A Subclass Analysis of Inductive Bias in Neural Morphology
Quick Answer
This study reveals that a rare irregular subtype in Japanese past-tense verb inflection accounts for significant model errors in neural morphological generation systems.
Quick Take
This study reveals that a rare irregular subtype in Japanese past-tense verb inflection accounts for significant model errors in neural morphological generation systems. By removing this specific subtype, generalization improves more than when all irregular verbs are excluded, highlighting the need for detailed subclass analysis in morphological evaluations.
Key Points
- Rare irregular subtype (<1% of data) causes disproportionate model errors.
- Removing this subtype improves generalization more than excluding all irregular verbs.
- Error concentration linked to low-frequency morphological patterns and gemination.
- Study suggests finer-grained subclass analysis for morphological evaluation.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Neural morphological generation systems often achieve high aggregate accuracy on benchmark datasets, yet such performance can conceal systematic errors concentrated in rare morphological subclasses. We examine Japanese past-tense verb inflection and show that a very small, structurally specific irregular subtype (<1% of data) accounts for a disproportionate share of model errors. Controlled ablation experiments demonstrate that removing this subtype yields larger improvements in generalization than removing all irregular verbs, indicating that not all irregularity contributes equally to model instability. These findings suggest that error concentration is driven by the interaction between extreme low-frequency morphological patterns and specific morphophonological processes, particularly gemination. We argue that morphological evaluation should incorporate finer-grained subclass analysis beyond standard conjugation categories.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2605.20558 [cs.CL] |
| (or arXiv:2605.20558v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2605.20558 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Wen Zhang [view email]
[v1]
Tue, 19 May 2026 23:18:47 UTC (33 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Quantifying Prior Dominance in Systems
The study introduces the Normalized Context Utilization (NCU) metric to evaluate Retrieval-Augmented Generation (RAG) systems, revealing that Small Language Models (SLMs) outperform larger models in factual extraction. The findings indicate that traditional scaling laws yield diminishing returns, with a commercial API frequently failing against adversarial evidence due to systemic confidence collapse.