The CUED's Grammatical Error Correction Systems for BEA-2019
This work addresses grammatical error correction for language learners, but it is incremental as it builds on prior methods and shared task participation.
The authors tackled grammatical error correction by submitting two systems to the BEA-2019 Shared Task: a low-resource system using finite state transducers with neural language models and a restricted neural system combining language and machine translation models with back-translation and fine-tuning.
We describe two entries from the Cambridge University Engineering Department to the BEA 2019 Shared Task on grammatical error correction. Our submission to the low-resource track is based on prior work on using finite state transducers together with strong neural language models. Our system for the restricted track is a purely neural system consisting of neural language models and neural machine translation models trained with back-translation and a combination of checkpoint averaging and fine-tuning -- without the help of any additional tools like spell checkers. The latter system has been used inside a separate system combination entry in cooperation with the Cambridge University Computer Lab.