Entity Classification on CoNLL04 (test)
93.8F1 ScoreBiLSTM-CRF with deep biaffine attention
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BiLSTM-CRF with deep biaffine attentionSetup=Setup 2 (Gold entity boundaries), Data Split=64/16/202018.12 | 93.8 | — | — | |
| Bekoulis et al. (2018)Setup=Setup 2 (Gold entity boundaries), Citation ID=[5]2018.12 | 93.3 | — | — | |
| Bekoulis et al. (2018)Setup=Setup 2 (Gold entity boundaries), Citation ID=[4]2018.12 | 93 | — | — | |
| Gupta et al. (2016)Setup=Setup 2 (Gold entity boundaries), Extra Features=true, Data Split=80/0/202018.12 | 92.4 | — | — | |
| Miwa and Sasaki (2014)Setup=Setup 2 (Gold entity boundaries), Extra Features=true, Data Split=80/0/202018.12 | 92.3 | — | — | |
| TriMFEvaluation protocol=strict2021.01 | 90.3 | 90.26 | 90.34 | |
| SpERTEvaluation protocol=strict2021.01 | 88.94 | 88.25 | 89.64 | |
| MRC4ERE++Evaluation protocol=strict2021.01 | 88.9 | 89.3 | 88.5 | |
| Gupta et al. (2016)Setup=Setup 2 (Gold entity boundaries), Data Split=80/0/202018.12 | 88.8 | — | — | |
| Multi-turn QAquestion type=Natural2019.05 | 87.8 | 89 | 86.6 | |
| Multi-turn QAEvaluation protocol=strict2021.01 | 87.8 | 89 | 86.6 | |
| Zhang et al. (2017)2019.05 | 85.6 | — | — | |
| Replicating Multihead with ATAvg. Epoch Train Time=614 sec, Avg. Test Time=34 sec2019.05 | 85.07 | 84.36 | 85.8 | |
| Relation-MetricAvg. Epoch Train Time=101 sec, Avg. Test Time=4.5 sec2019.05 | 84.57 | 84.46 | 84.67 | |
| Multihead2019.05 | 83.9 | 83.75 | 84.06 | |
| Multi-head + ATEvaluation protocol=strict2021.01 | 83.9 | — | — | |
| Multihead with AT2019.05 | 83.61 | — | — | |
| Bekoulis et al. (2018)2019.05 | 83.6 | — | — | |
| Adel and Schütze (2017)Setup=Setup 2 (Gold entity boundaries)2018.12 | 82.1 | — | — | |
| Table Representation2019.05 | 80.7 | 81.2 | 80.2 | |
| Miwa and Sasaki (2014)2019.05 | 80.7 | — | — |