A Sub-Character Architecture for Korean Language Processing

TL;DR

This paper presents a novel sub-character architecture for Korean language processing that decomposes characters into phonetic units called jamo, improving accuracy and reducing data sparsity in dependency parsing.

Contribution

The paper introduces a new sub-character architecture leveraging jamo decomposition, enhancing Korean NLP tasks by capturing syntactic and semantic information more effectively.

Findings

Reduced observation space to 1.6% of original

Significant accuracy improvements in dependency parsing

Effective alleviation of data sparsity issues

Abstract

We introduce a novel sub-character architecture that exploits a unique compositional structure of the Korean language. Our method decomposes each character into a small set of primitive phonetic units called jamo letters from which character- and word-level representations are induced. The jamo letters divulge syntactic and semantic information that is difficult to access with conventional character-level units. They greatly alleviate the data sparsity problem, reducing the observation space to 1.6% of the original while increasing accuracy in our experiments. We apply our architecture to dependency parsing and achieve dramatic improvement over strong lexical baselines.