Grammatical Error Correction On Conll 2014
评估指标
F0.5
Precision
Recall
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | F0.5 | Precision | Recall |
|---|---|---|---|
| gector-grammatical-error-correction-tag-not | 66.5 | 78.2 | 41.5 |
| neural-quality-estimation-of-grammatical | 56.52 | - | - |
| approaching-neural-grammatical-error | 55.8 | - | - |
| lm-critic-language-models-for-unsupervised | 65.8 | - | - |
| improving-seq2seq-grammatical-error | 69.6 | 79.2 | 46.8 |
| parallel-iterative-edit-models-for-local | 61.2 | - | - |
| unsupervised-grammatical-error-correction | 69.6 | 75.0 | 53.8 |
| frustratingly-easy-system-combination-for | 69.51 | 81.48 | 43.78 |
| an-empirical-study-of-incorporating-pseudo | 65.0 | - | - |
| near-human-level-performance-in-grammatical | 56.25 | - | - |
| stronger-baselines-for-grammatical-error | 63.0 | 69.9 | 45.1 |
| pillars-of-grammatical-error-correction | 71.8 | 83.7 | 45.7 |
| syngec-syntax-enhanced-grammatical-error | 67.6 | 74.7 | 49.0 |
| gector-grammatical-error-correction-tag-not | 65.3 | 77.5 | 40.1 |
| system-combination-via-quality-estimation-for | 71.12 | 79.6 | 49.86 |
| improving-grammatical-error-correction-via | 61.15 | 71.57 | 38.65 |
| a-simple-recipe-for-multilingual-grammatical | 68.87 | - | - |
| a-multilayer-convolutional-encoder-decoder | 54.79 | - | - |
| pillars-of-grammatical-error-correction | 72.8 | 83.9 | 47.5 |
| parallel-iterative-edit-models-for-local | 59.7 | - | - |
| encoder-decoder-models-can-benefit-from-pre | 65.2 | - | - |
| neural-quality-estimation-with-multiple | 63.7 | - | - |
| efficient-and-interpretable-grammatical-error | 67.79 | 74.29 | 50.21 |