Grammatical Error Detection On Conll 2014 A1
Metriken
F0.5
Ergebnisse
Leistungsergebnisse verschiedener Modelle zu diesem Benchmark
| Paper Title | ||
|---|---|---|
| VERNet | 54.3 | Neural Quality Estimation with Multiple Hypotheses for Grammatical Error Correction |
| Bi-LSTM + POS (unrestricted data) | 36.1 | Auxiliary Objectives for Neural Error Detection Models |
| Bi-LSTM (unrestricted data) | 34.3 | Compositional Sequence Labeling Models for Error Detection in Learner Writing |
| BiLSTM-JOINT (trained on FCE) | 22.14 | Jointly Learning to Label Sentences and Tokens |
| Ann+PAT+MT | 21.87 | Artificial Error Generation with Machine Translation and Syntactic Patterns |
| Bi-LSTM + LMcost (trained on FCE) | 17.86 | Semi-supervised Multitask Learning for Sequence Labeling |
| Bi-LSTM + POS (trained on FCE) | 17.5 | Auxiliary Objectives for Neural Error Detection Models |
| Bi-LSTM (trained on FCE) | 16.4 | Compositional Sequence Labeling Models for Error Detection in Learner Writing |
0 of 8 row(s) selected.