Neural Chinese Word Segmentation as Sequence to Sequence Translation

TL;DR

This paper introduces a novel sequence-to-sequence neural model for Chinese word segmentation that captures global context and can be extended to joint tasks like spelling correction, achieving competitive results.

Contribution

It proposes a sequence-to-sequence approach with attention for CWS, enabling global context modeling and multi-task learning, which is a departure from traditional local feature-based methods.

Findings

Achieved competitive performance on benchmark datasets

Successfully applied to joint CWS and spelling correction

Demonstrated effectiveness of global context modeling

Abstract

Recently, Chinese word segmentation (CWS) methods using neural networks have made impressive progress. Most of them regard the CWS as a sequence labeling problem which construct models based on local features rather than considering global information of input sequence. In this paper, we cast the CWS as a sequence translation problem and propose a novel sequence-to-sequence CWS model with an attention-based encoder-decoder framework. The model captures the global information from the input and directly outputs the segmented sequence. It can also tackle other NLP tasks with CWS jointly in an end-to-end mode. Experiments on Weibo, PKU and MSRA benchmark datasets show that our approach has achieved competitive performances compared with state-of-the-art methods. Meanwhile, we successfully applied our proposed model to jointly learning CWS and Chinese spelling correction, which demonstrates its applicability of multi-task fusion.