Research on proofreading methods of Chinese semantic knowledge

Inhalt:
With the rapid increase of the trend of electronic text, it is more urgent to improve the performance of Chinese text auto- proofreading models. Due to the effective application of deep learning technology, the research on Chinese spelling proofreading and Chinese grammar proofreading has been relatively perfect, but the research on Chinese semantic knowledge proofreading methods is still less, resulting in semantic knowledge proofreading is a weak link in Chinese text proofreading. In this paper, the semantic knowledge proofreading data set is constructed by relation extraction and other techniques, aiming to make the proofreading model can further improve the knowledge reasoning ability in combination with the context. The experimental results show that the method proposed in this paper further improves the knowledge reasoning ability of the model, and can make better use of the semantic knowledge contained in the pretrained model, which provides a new idea for the following research.