Improving Korean NLP Tasks with Linguistically Informed Subword Tokenization and Sub-character Decomposition

TL;DR

This paper presents a linguistically informed subword tokenization method for Korean that improves language model performance by incorporating sub-character decomposition and morpheme awareness, especially enhancing syntactic tasks.

Contribution

The paper introduces a novel morpheme-aware subword tokenization approach utilizing sub-character decomposition tailored for Korean, balancing linguistic accuracy and computational efficiency.

Findings

Improved performance on Korean NLP tasks, notably NIKL-CoLA.

Enhanced syntactic and semantic capabilities in language models.

Potential for further improvements with more linguistic insights.

Abstract

We introduce a morpheme-aware subword tokenization method that utilizes sub-character decomposition to address the challenges of applying Byte Pair Encoding (BPE) to Korean, a language characterized by its rich morphology and unique writing system. Our approach balances linguistic accuracy with computational efficiency in Pre-trained Language Models (PLMs). Our evaluations show that this technique achieves good performances overall, notably improving results in the syntactic task of NIKL-CoLA. This suggests that integrating morpheme type information can enhance language models' syntactic and semantic capabilities, indicating that adopting more linguistic insights can further improve performance beyond standard morphological analysis.