Glyce: Glyph-vectors for Chinese Character Representations

TL;DR

Glyce introduces glyph-vectors for Chinese characters by leveraging historical scripts, specialized CNNs, and multi-task learning, significantly improving performance across various Chinese NLP tasks.

Contribution

The paper presents a novel glyph-based representation method for Chinese characters, incorporating historical scripts and tailored CNNs, achieving state-of-the-art results in multiple NLP tasks.

Findings

Outperforms ID-based models in Chinese NLP tasks

Achieves new state-of-the-art results in NER, CWS, POS, and text classification

Demonstrates the effectiveness of historical script data and multi-task learning

Abstract

It is intuitive that NLP tasks for logographic languages like Chinese should benefit from the use of the glyph information in those languages. However, due to the lack of rich pictographic evidence in glyphs and the weak generalization ability of standard computer vision models on character data, an effective way to utilize the glyph information remains to be found. In this paper, we address this gap by presenting Glyce, the glyph-vectors for Chinese character representations. We make three major innovations: (1) We use historical Chinese scripts (e.g., bronzeware script, seal script, traditional Chinese, etc) to enrich the pictographic evidence in characters; (2) We design CNN structures (called tianzege-CNN) tailored to Chinese character image processing; and (3) We use image-classification as an auxiliary task in a multi-task learning setup to increase the model's ability to generalize. We show that glyph-based models are able to consistently outperform word/char ID-based models in a wide range of Chinese NLP tasks. We are able to set new state-of-the-art results for a variety of Chinese NLP tasks, including tagging (NER, CWS, POS), sentence pair classification, single sentence classification tasks, dependency parsing, and semantic role labeling. For example, the proposed model achieves an F1 score of 80.6 on the OntoNotes dataset of NER, +1.5 over BERT; it achieves an almost perfect accuracy of 99.8\% on the Fudan corpus for text classification. Code found at https://github.com/ShannonAI/glyce.