Grammatical Error Correction on CoNLL 2014
65.2F0.5Kaneko et al.
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Kaneko et al.Configuration=Ensemble2020.05 | 65.2 | 72.6 | 46.4 | |
| Kiyono et al.Configuration=Ensemble2020.05 | 64.7 | 73.3 | 44.2 | |
| BART-basedConfiguration=Ensemble2020.05 | 63 | 69.9 | 45.1 | |
| Kaneko et al.Configuration=Single2020.05 | 62.6 | 69.2 | 45.6 | |
| BART-basedConfiguration=Single2020.05 | 62.6 | 69.3 | 45 | |
| Base + FB learningfluency boost learning=true2018.07 | 61.34 | — | — | |
| Base + FB learningtraining=fluency boost learning2018.07 | 61.34 | 74.12 | 36.3 | |
| Kiyono et al.Configuration=Single2020.05 | 61.3 | 67.9 | 44.1 | |
| ELECTRA-VERNetTop-K=3, Backbone=ELECTRA2021.05 | 60.35 | 67.51 | 42.38 | |
| ELECTRA-VERNetTop-K=4, Backbone=ELECTRA2021.05 | 60.25 | 67.48 | 42.19 | |
| ELECTRA-VERNetTop-K=2, Backbone=ELECTRA2021.05 | 60.22 | 66.95 | 42.97 | |
| ELECTRA-VERNetTop-K=5, Backbone=ELECTRA2021.05 | 60.01 | 67.15 | 42.1 | |
| Base + FB learning and inferencefluency boost learning=true, fluency boost inference=true2018.07 | 60 | — | — | |
| Base + FB learning and inferencetraining=fluency boost learning, inference=fluency boost inference2018.07 | 60 | 68.45 | 40.18 | |
| BERT-VERNetTop-K=4, Backbone=BERT2021.05 | 59.21 | 66.65 | 40.94 | |
| BERT-VERNetTop-K=2, Backbone=BERT2021.05 | 58.98 | 65.62 | 41.98 | |
| BERT-VERNetTop-K=5, Backbone=BERT2021.05 | 58.98 | 66.41 | 40.74 | |
| BERT-VERNetTop-K=3, Backbone=BERT2021.05 | 58.86 | 65.92 | 41.22 | |
| Basic GEC2021.05 | 58.69 | 64.26 | 43.59 | |
| Base convolutional seq2seq2018.07 | 57.95 | — | — | |
| Base convolutional seq2seq2018.07 | 57.95 | 72.52 | 32.13 | |
| SMT-NMT hybrid2018.07 | 56.25 | — | — | |
| SMT-NMT hybrid2018.07 | 56.25 | 66.77 | 34.49 | |
| Adapted-transformer2018.07 | 55.8 | — | — | |
| Adapted-transformer2018.07 | 55.8 | 61.9 | 40.2 | |
| NUS182018.07 | 54.79 | — | — | |
| NUS182018.07 | 54.79 | 65.49 | 33.14 | |
| NUS172018.07 | 53.14 | — | — | |
| NUS172018.07 | 53.14 | 62.74 | 32.96 | |
| CAMB172018.07 | 51.08 | — | — | |
| AMU162018.07 | 49.49 | — | — | |
| Back-CNN-seq2seq2018.07 | 49 | — | — | |
| VT162018.07 | 47.4 | — | — | |
| Nested-RNN-seq2seq2018.07 | 45.15 | — | — | |
| NUS162018.07 | 44.27 | — | — | |
| CAMB162018.07 | 39.9 | — | — | |
| CAMB142018.07 | 37.33 | — | — | |
| CUUI2018.07 | 36.79 | — | — | |
| AMU142018.07 | 35.01 | — | — |