Entity recognition on SCIERC (test)
70.33F1 ScoreSpERT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SpERTEvaluation protocol=boundary, Backbone=SciBERT2021.01 | 70.33 | 70.87 | 69.79 | — | |
| TriMFEncoder=SciBERT, Representation Type=T-Concat2021.09 | 70.2 | — | — | — | |
| TriMFEvaluation protocol=boundary, Backbone=SciBERT2021.01 | 70.17 | 70.18 | 70.17 | — | |
| PL-MarkerEncoder=SciBERT, Representation Type=Solid & Levitated Marker, Cross-sentence information=true2021.09 | 69.9 | — | — | — | |
| PURE-FEncoder=SciBERT, Representation Type=Solid Marker2021.09 | 68.9 | — | — | — | |
| UniREEncoder=SciBERT, Representation Type=T-Concat2021.09 | 68.1 | — | — | — | |
| DYGIE++Backbone=BERT-Base2019.09 | 67.5 | — | — | — | |
| DYGIE++Evaluation protocol=boundary2021.01 | 67.5 | — | — | — | |
| DYGIE2019.04 | 65.2 | — | — | — | |
| DYGIEBackbone=ELMo2019.09 | 65.2 | — | — | — | |
| DyGIEEvaluation protocol=boundary2021.01 | 65.2 | — | — | — | |
| DYGIEEncoder=ELMo, Representation Type=T-Concat2021.09 | 65.2 | — | — | — | |
| SCIIEmulti-task learning=true2018.08 | 64.2 | 67.2 | 61.5 | — | |
| Luan et al.2019.04 | 64.2 | — | — | — | |
| SciIEEvaluation protocol=boundary2021.01 | 64.2 | 67.2 | 61.5 | — | |
| E2E Rel+ELMOfeature=ELMO2018.08 | 63.7 | 63.5 | 63.9 | — | |
| LSTM+CRF+ELMOfeature=ELMO2018.08 | 63.5 | 63.8 | 63.2 | — | |
| LSTM+CRF2018.08 | 62 | 62.9 | 61.1 | — | |
| E2E Rel2018.08 | 61.2 | 60.6 | 61.9 | — | |
| E2E Rel(Pipeline)setting=pipeline2018.08 | 61 | 60.8 | 61.2 | — | |
| BERT-BaseEvaluation Protocol=Frozen2019.03 | — | — | — | 63.58 | |
| BERT-BaseEvaluation Protocol=Finetune2019.03 | — | — | — | 65.24 | |
| SCIBERTEvaluation Protocol=Frozen2019.03 | — | — | — | 65.77 | |
| SCIBERTEvaluation Protocol=Finetune2019.03 | — | — | — | 67.57 | |
| SOTA (Luan et al., 2018)2019.03 | — | — | — | 64.2 |