ISCA Archive SLaTE 2017
ISCA Archive SLaTE 2017

Deep Context Model for Grammatical Error Correction

Chuan Wang, Ruobing Li, Hui Lin

In this paper, we propose a deep context model based on recurrent neural networks (RNN) for grammatical error correction. For a specific error type, we treat the error correction task as a classification problem where the grammatical context representation is learnt from native text data that are largely available. Compared with traditional classifier methods, our model does not require sophisticated feature engineering which usually requires linguistic knowledge and may not cover all context patterns.Experiments on CoNLL-2014 shared task show that our approach significantly outperforms the state-of-the-art classifier and machine translation approaches for grammatical error correction.


doi: 10.21437/SLaTE.2017-29

Cite as: Wang, C., Li, R., Lin, H. (2017) Deep Context Model for Grammatical Error Correction. Proc. 7th ISCA Workshop on Speech and Language Technology in Education (SLaTE 2017), 167-171, doi: 10.21437/SLaTE.2017-29

@inproceedings{wang17_slate,
  author={Chuan Wang and Ruobing Li and Hui Lin},
  title={{Deep Context Model for Grammatical Error Correction}},
  year=2017,
  booktitle={Proc. 7th ISCA Workshop on Speech and Language Technology in Education (SLaTE 2017)},
  pages={167--171},
  doi={10.21437/SLaTE.2017-29}
}