Semantic Role Labeling on CoNLL 2005 (WSJ (in-domain))
88.93F1 ScoreSynG2G-Tr
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SynG2G-TrSA=true, Evaluation Setting=given predicate, Pre-training=true2021.04 | 88.93 | 89.11 | 88.74 | |
| Jia et al. (2022)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 88.25 | — | — | |
| Jindal et al. (2020)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 87.93 | 87.7 | 88.15 | |
| Zhang et al. (2021)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 87.9 | 88.7 | 88 | |
| Li et al. (2019)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 87.7 | 87.9 | 87.5 | |
| Ouchi et al. (2018)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 87.6 | 88.2 | 87 | |
| SynG2G-TrSA=true, Evaluation Setting=end-to-end, Pre-training=true2021.04 | 87.57 | 86.86 | 88.3 | |
| He et al. (2018a)SA=false, Evaluation Setting=given predicate, Pre-training=true2021.04 | 87.4 | — | — | |
| Strubell et al. (2018)†SA=true, Evaluation Setting=end-to-end, Pre-training=true2021.04 | 86.9 | 87.13 | 86.67 | |
| SynG2G-Tr (w/o BERT)SA=true, Evaluation Setting=given predicate, Pre-training=false2021.04 | 86.5 | 86.46 | 86.56 | |
| Li et al. (2019)SA=false, Evaluation Setting=end-to-end, Pre-training=true2021.04 | 86.3 | 85.2 | 87.5 | |
| Strubell et al. (2018)†SA=true, Evaluation Setting=given predicate, Pre-training=false2021.04 | 86.04 | 86.02 | 86.05 | |
| He et al. (2018a)SA=false, Evaluation Setting=end-to-end, Pre-training=true2021.04 | 86 | 84.8 | 87.2 | |
| SynG2G-Tr (w/o BERT)SA=true, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 85.45 | 84.48 | 86.46 | |
| Xia et al. (2020)SA=true, Evaluation Setting=given predicate, Pre-training=false2021.04 | 85.06 | 85.12 | 85 | |
| Strubell et al. (2018)SA=true, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 84.99 | 85.53 | 84.45 | |
| Tan et al. (2017)SA=false, Evaluation Setting=given predicate, Pre-training=false2021.04 | 84.8 | 84.5 | 85.2 | |
| He et al. (2017)SA=false, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 84.6 | 85 | 84.3 | |
| Xia et al. (2020)SA=true, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 84.49 | 83.05 | 84.49 | |
| Xia et al. (2019)SA=true, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 84.1 | 84.3 | 83.8 | |
| He et al. (2018a)SA=false, Evaluation Setting=given predicate, Pre-training=false2021.04 | 83.9 | — | — | |
| Ouchi et al. (2018)SA=false, Evaluation Setting=given predicate, Pre-training=false2021.04 | 83.5 | 84.7 | 82.3 | |
| Li et al. (2019)SA=true, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 83 | — | — | |
| He et al. (2018a)SA=false, Evaluation Setting=end-to-end, Pre-training=false2021.04 | 82.5 | 81.2 | 83.9 |