Deep Text Corrector

Work in Progress

Introduction

This project aims to make a text corrector for English learners using deep neural net. This project is implemented in Pytorch

• First trial: seq2seq w/ attention + nucle 2.3 dataset
• Next..
  • more data, data augmentations, ...
  • tweeking nets like Bytenet or Transformer
  • beam search

Datasets

CoNLL-2013 Shared Task: Grammatical Error Correction Overview Paper Datasets Participant Papers

CoNLL-2014 Shared Task: Grammatical Error Correction Overview Paper [Datasets] Needs a license form.

References

https://github.com/atpaino/deep-text-corrector