Relation Extraction on ADE (test)
90Macro F1BioBERT
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BioBERTArchitecture=Accuracy-Optimized BioBERT2021.12 | 90 | — | — | — | — | |
| FCNNArchitecture=Speed-Optimized FCNN2021.12 | 89.2 | — | — | — | — | |
| Oursvalidation=10-fold cross-validation, classifier=KNN classifier, model=CLDR2021.09 | 86.5 | — | — | — | — | |
| Crone 20202021.12 | 83.7 | — | — | — | — | |
| Zhao et al., 2020validation=10-fold cross-validation2021.09 | 81.14 | — | — | — | — | |
| Wang and Lu, 2020validation=10-fold cross-validation2021.09 | 80.1 | — | — | — | — | |
| Oursvalidation=10-fold cross-validation, model=CLDR2021.09 | 79.97 | — | — | — | — | |
| Eberts and Ulges, 2020validation=10-fold cross-validation2021.09 | 79.24 | — | — | — | — | |
| Tran and Kavuluru, 2019validation=10-fold cross-validation2021.09 | 77.29 | — | — | — | — | |
| Bekoulis et al., 2018avalidation=10-fold cross-validation2021.09 | 75.52 | — | — | — | — | |
| Bekoulis et al., 2018bvalidation=10-fold cross-validation2021.09 | 74.58 | — | — | — | — | |
| Li et al., 2017validation=10-fold cross-validation2021.09 | 71.4 | — | — | — | — | |
| Li et al., 2016validation=10-fold cross-validation2021.09 | 63.4 | — | — | — | — | |
| Bekoulis et al.Averaging=macro-averaged2020.10 | — | — | 74.6 | — | — | |
| Bekoulis et al.Averaging=macro-averaged2020.10 | — | — | 75.5 | — | — | |
| BioBERT Relation Extraction modelArchitecture=BioBERT2022.01 | — | 80.1 | — | 81.31 | 79.03 | |
| Eberts and UlgesAveraging=macro-averaged2020.10 | — | — | 79.2 | — | — | |
| FCNN Relation Extraction modelArchitecture=FCNN2022.01 | — | 74.7 | — | 69.11 | 86.22 | |
| Li et al.Averaging=macro-averaged2020.10 | — | — | 63.4 | — | — | |
| Li et al.Averaging=macro-averaged2020.10 | — | — | 71.4 | — | — | |
| Multi-head + ATAveraging method=macro-average2021.01 | — | 75.52 | — | — | — | |
| SOTA2022.01 | — | 83.74 | — | — | — | |
| SpERTEvaluation protocol=strict, Averaging method=macro-average2021.01 | — | 78.84 | — | 77.77 | 79.96 | |
| Table-Sequence EncodersAveraging=macro-averaged2020.10 | — | 80.1 | 80.1 | — | — | |
| Tran and KavuluruAveraging=macro-averaged2020.10 | — | — | 77.3 | — | — | |
| TriMFEvaluation protocol=boundary, Averaging method=macro-average2021.01 | — | 80.66 | — | 74.22 | 88.48 |