Enhancing Korean Dependency Parsing with Morphosyntactic Features

2025-03-26

AI Summary

Korean’s rich inflectional morphology and free word order cause a disconnect between syntactic and morphological modeling in existing dependency parsing frameworks, leading to annotation inconsistencies and reduced parsing accuracy. To address this, we propose UniDive—the first unified framework jointly modeling Universal Dependencies (UD) and Universal Morphology (UniMorph) annotations for Korean. UniDive explicitly integrates morphological features into dependency parsing via multi-task learning and is compatible with both encoder-only (e.g., BERT) and decoder-only (e.g., LLaMA) architectures. It enables, for the first time, end-to-end co-modeling of Korean syntactic dependencies and fine-grained morphological features. Experiments demonstrate significant improvements in overall UAS and LAS, with substantial gains on morphology-sensitive relations—particularly case marking and tense—thereby empirically validating the critical role of morphological information in syntactic relation disambiguation.

Technology Category

• Natural Language Processing: Lexical Semantics and Morphology
• Machine Learning: Large Multimodal Models (LMMs)
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

This paper introduces UniDive for Korean, an integrated framework that bridges Universal Dependencies (UD) and Universal Morphology (UniMorph) to enhance the representation and processing of Korean {morphosyntax}. Korean's rich inflectional morphology and flexible word order pose challenges for existing frameworks, which often treat morphology and syntax separately, leading to inconsistencies in linguistic analysis. UniDive unifies syntactic and morphological annotations by preserving syntactic dependencies while incorporating UniMorph-derived features, improving consistency in annotation. We construct an integrated dataset and apply it to dependency parsing, demonstrating that enriched morphosyntactic features enhance parsing accuracy, particularly in distinguishing grammatical relations influenced by morphology. Our experiments, conducted with both encoder-only and decoder-only models, confirm that explicit morphological information contributes to more accurate syntactic analysis.

Problem

• Enhancing Korean dependency parsing accuracy
• Unifying syntactic and morphological annotations
• Addressing inconsistencies in linguistic analysis

Innovation

• UniDive integrates UD and UniMorph for Korean
• Combines syntactic dependencies with UniMorph features
• Enhances parsing accuracy using morphosyntactic features