The Impact of Visual Information in Chinese Characters: Evaluating Large Models' Ability to Recognize and Utilize Radicals

TL;DR

This paper evaluates whether large language and vision-language models can recognize and utilize visual features of Chinese characters, such as radicals, and demonstrates that prompting with radicals can improve language processing tasks.

Contribution

It establishes a benchmark for assessing models' understanding of Chinese character visuals and shows that prompting with radicals can enhance Chinese language processing performance.

Findings

Models show limited understanding of visual features in Chinese characters.

Providing radical information improves Part-Of-Speech tagging accuracy.

Visual features are underutilized by current large models.

Abstract

The glyphic writing system of Chinese incorporates information-rich visual features in each character, such as radicals that provide hints about meaning or pronunciation. However, there has been no investigation into whether contemporary Large Language Models (LLMs) and Vision-Language Models (VLMs) can harness these sub-character features in Chinese through prompting. In this study, we establish a benchmark to evaluate LLMs' and VLMs' understanding of visual elements in Chinese characters, including radicals, composition structures, strokes, and stroke counts. Our results reveal that models surprisingly exhibit some, but still limited, knowledge of the visual information, regardless of whether images of characters are provided. To incite models' ability to use radicals, we further experiment with incorporating radicals into the prompts for Chinese language processing (CLP) tasks. We observe consistent improvement in Part-Of-Speech tagging when providing additional information about radicals, suggesting the potential to enhance CLP by integrating sub-character information.