CUNI System for the Building Educational Applications 2019 Shared Task: Grammatical Error Correction
This work addresses grammatical error correction for educational applications, presenting incremental improvements in shared task performance.
The paper tackles grammatical error correction by developing Transformer-based NMT systems with enhancements like dropout and iterative correction, achieving F0.5 scores of 59.39 in the Restricted Track, 44.13 in the Low-Resource Track, and 64.55 in the Unrestricted Track.
In this paper, we describe our systems submitted to the Building Educational Applications (BEA) 2019 Shared Task (Bryant et al., 2019). We participated in all three tracks. Our models are NMT systems based on the Transformer model, which we improve by incorporating several enhancements: applying dropout to whole source and target words, weighting target subwords, averaging model checkpoints, and using the trained model iteratively for correcting the intermediate translations. The system in the Restricted Track is trained on the provided corpora with oversampled "cleaner" sentences and reaches 59.39 F0.5 score on the test set. The system in the Low-Resource Track is trained from Wikipedia revision histories and reaches 44.13 F0.5 score. Finally, we finetune the system from the Low-Resource Track on restricted data and achieve 64.55 F0.5 score, placing third in the Unrestricted Track.