Entity Extraction on ACE05 (test)
91.1F1 ScorePL-Marker
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PL-MarkerEncoder=ALBERT_XXL, Representation Type=Solid & Levitated Marker2021.09 | 91.1 | — | — | |
| PURE-FEncoder=ALBERT_XXL, Representation Type=Solid Marker2021.09 | 90.9 | — | — | |
| UniREEncoder=ALBERT_XXL, Representation Type=T-Concat2021.09 | 90.2 | — | — | |
| PURE-FEncoder=BERT_BASE, Representation Type=Solid Marker2021.09 | 90.1 | — | — | |
| PL-MarkerEncoder=BERT_BASE, Representation Type=Solid & Levitated Marker, Cross-sentence information=true2021.09 | 89.8 | — | — | |
| TableSeqEncoder=ALBERT_XXL, Representation Type=T-Concat2021.09 | 89.5 | — | — | |
| OneIEEncoder=BERT_L, Representation Type=T-Concat2021.09 | 88.8 | — | — | |
| UniREEncoder=BERT_BASE, Representation Type=T-Concat2021.09 | 88.8 | — | — | |
| DYGIE++Backbone=BERT-Base2019.09 | 88.6 | — | — | |
| DYGIE++Evaluation protocol=boundary2021.01 | 88.6 | — | — | |
| DyGIE++ (Wadden et al., 2019)Evaluation Criteria=ACE05♦2020.10 | 88.6 | — | — | |
| DYGIE++Encoder=BERT_BASE, Representation Type=T-Concat2021.09 | 88.6 | — | — | |
| DYGIE2019.04 | 88.4 | — | — | |
| DYGIEBackbone=ELMo2019.09 | 88.4 | — | — | |
| DyGIEEvaluation protocol=boundary2021.01 | 88.4 | — | — | |
| DYGIEEncoder=ELMo, Representation Type=T-Concat2021.09 | 88.4 | — | — | |
| TablERTEvaluation Criteria=ACE05♦2020.10 | 88 | 87.8 | 88.2 | |
| TablERTEvaluation Criteria=ACE05♠2020.10 | 88 | 87.8 | 88.2 | |
| TriMFEvaluation protocol=boundary2021.01 | 87.61 | 87.67 | 87.54 | |
| TriMFEvaluation protocol=strict2021.01 | 87.61 | 87.67 | 87.54 | |
| TriMFEncoder=BERT_BASE, Representation Type=T-Concat2021.09 | 87.6 | — | — | |
| Sanh et al.2019.04 | 87.5 | — | — | |
| Dixit and Al-Onaizan (2019)Evaluation Criteria=ACE05♦2020.10 | 86 | 85.9 | 86.1 | |
| MRC4ERE++Evaluation protocol=strict2021.01 | 85.5 | 85.9 | 85.2 | |
| Multi-turn QA2019.05 | 84.8 | 84.7 | 84.9 | |
| Multi-turn QAEvaluation protocol=strict2021.01 | 84.8 | 84.7 | 84.9 | |
| Multi-turn QA (Li et al., 2019)Evaluation Criteria=ACE05♠2020.10 | 84.8 | 84.7 | 84.9 | |
| Multi-turn QAEncoder=BERT_L2021.09 | 84.8 | — | — | |
| Zhang et al.2019.04 | 83.6 | — | — | |
| Sun et al.2019.05 | 83.6 | 83.9 | 83.2 | |
| Zhang et al. (2017)Evaluation Criteria=ACE05♠2020.10 | 83.6 | — | — | |
| MRT (Sun et al., 2018)Evaluation Criteria=ACE05♠2020.10 | 83.6 | 83.9 | 83.2 | |
| Zhang et al.2019.05 | 83.5 | — | — | |
| SPTree2016.01 | 83.4 | 82.9 | 83.9 | |
| Miwa and Bansal2019.04 | 83.4 | — | — | |
| Miwa and Bansal2019.05 | 83.4 | 82.9 | 83.9 | |
| SPTree (Miwa and Bansal, 2016)Evaluation Criteria=ACE05♠2020.10 | 83.4 | 82.9 | 83.9 | |
| SPtreeEncoder=LSTM, Representation Type=T-Concat2021.09 | 83.4 | — | — | |
| DyGIEEmbeddings=Language Model (LM), Additional training data=OntoNotes2019.06 | 82.9 | — | — | |
| Katiyar and Cardie2019.05 | 82.6 | 84 | 81.3 | |
| Merge and Label [BERT]Embeddings=BERT, POS tags usage=false2019.06 | 82.4 | 82.7 | 82.1 | |
| Li and Ji2016.01 | 80.8 | 85.2 | 76.9 | |
| Li and Ji2019.05 | 80.8 | 85.2 | 76.9 | |
| Li and Ji (2014)Evaluation Criteria=ACE05♦2020.10 | 80.8 | 85.2 | 76.9 | |
| Li and Ji (2014)Evaluation Criteria=ACE05♠2020.10 | 80.8 | 85.2 | 76.9 | |
| Li and JiEncoder=BERT_L2021.09 | 80.8 | — | — | |
| Merge and Label [ELMO]Embeddings=ELMO, POS tags usage=false2019.06 | 78.9 | 79.7 | 78 | |
| Merge and LabelEmbeddings=Glove, POS tags usage=false2019.06 | 74.6 | 75.1 | 74.1 | |
| Segm. hyp [POS]Embeddings=Glove, POS tags usage=true2019.06 | 74.5 | 76.8 | 72.3 | |
| LSTM + forest [POS]Embeddings=Glove, POS tags usage=true2019.06 | 73 | 74.5 | 71.5 | |
| BiLSTM-CRF stackedEmbeddings=Glove, POS tags usage=false2019.06 | 72.2 | 74.2 | 70.3 | |
| RNN + hypEmbeddings=Glove, POS tags usage=false2019.06 | 70.5 | 70.6 | 70.4 | |
| Multigraph + MSEmbeddings=Glove, POS tags usage=false2019.06 | 63.1 | 69.1 | 58.1 |