Sequence tagging + token-level transformations + two-stage fine-tuning (+XLNet)

Reported on 4 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing4 results

Grammatical Error CorrectiononCoNLL-2014 Shared Task
F0.5· 2020-05-26
65.3
best: 72.8 (Ensembles of best 7 models + GRECO + GTP-rerank)
GECToR -- Grammatical Error Correction: Tag, Not Rewrite arXiv:2005.12592
Grammatical Error CorrectiononCoNLL-2014 Shared Task
Precision· 2020-05-26
77.5
best: 83.9 (Ensembles of best 7 models + GRECO + GTP-rerank)
GECToR -- Grammatical Error Correction: Tag, Not Rewrite arXiv:2005.12592
Grammatical Error CorrectiononCoNLL-2014 Shared Task
Recall· 2020-05-26
40.1
best: 53.8 (Unsupervised GEC + cLang8)
GECToR -- Grammatical Error Correction: Tag, Not Rewrite arXiv:2005.12592
Grammatical Error CorrectiononBEA-2019 (test)
F0.5· 2020-05-26
72.4
best: 81.4 (Majority-voting ensemble on best 7 models)
GECToR -- Grammatical Error Correction: Tag, Not Rewrite arXiv:2005.12592