[arXiv]score: 0.07
Korean GEC Annotation Scheme Fixes Morpheme-Word Mismatch in L2 Corpora
June 1, 2026
A refined annotation pipeline converts the NIKL L2 Korean corpus from morpheme-level to word-level m2 edits under morphological constraints, and defines a Korean ERRANT-style scheme distinguishing functional morpheme, spelling, word boundary, and word order errors.
cs.CL
HOW THIS AFFECTS YOU
●
researcherDirectly useful if you work on Korean GEC or low-resource morphologically rich language evaluation — provides a reproducible annotation framework and corrected corpus.