Named Entity Recognition on MSRA (test)
96.24F1 ScoreMECT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MECTBackbone=BERT2021.07 | 96.24 | — | — | |
| W2NER2021.12 | 96.1 | 96.12 | 96.08 | |
| Li et al. (2020b)2021.12 | 96.09 | — | — | |
| Clean dataMethod Category=Upper Bound2023.05 | 95.53 | — | — | |
| SoftLexicon (LSTM) + BERTArchitecture=SoftLexicon, Sequence modeling=LSTM, Backbone=BERT2019.08 | 95.42 | 95.75 | 95.1 | |
| SoftLexicon (LSTM) + BERT2019.08 | 95.42 | 95.75 | 95.1 | |
| Ma et al. (2020)2021.12 | 95.42 | 95.75 | 95.1 | |
| CPLLType=Ours2023.05 | 95.42 | — | — | |
| LW lossMethod Category=PLL2023.05 | 95.35 | — | — | |
| BERT2021.07 | 94.95 | — | — | |
| BERT + LSTM + CRFArchitecture=BERT + LSTM + CRF2019.08 | 94.83 | 95.06 | 94.61 | |
| Entity-level VotingMethod Category=Voting2023.05 | 94.78 | — | — | |
| Token-level VotingMethod Category=Voting2023.05 | 94.68 | — | — | |
| MECT2021.07 | 94.32 | 94.55 | 94.09 | |
| Baseline2021.07 | 94.12 | — | — | |
| SoftLexicon (LSTM) + bicharArchitecture=SoftLexicon, Sequence modeling=LSTM, Features=bichar2019.08 | 94.06 | 94.73 | 93.4 | |
| SoftLexiconArchitecture=LSTM, Features=bichar2021.07 | 94.06 | 94.73 | 93.4 | |
| BERT-TaggerArchitecture=BERT2019.08 | 93.76 | 93.4 | 94.12 | |
| LR-CNNArchitecture=LR-CNN2019.08 | 93.71 | 94.5 | 92.93 | |
| LR-CNN2021.07 | 93.71 | 94.5 | 92.93 | |
| Gui et al. (2019)2021.12 | 93.71 | 94.5 | 92.93 | |
| SoftLexicon (LSTM)Architecture=SoftLexicon, Sequence modeling=LSTM2019.08 | 93.66 | 94.63 | 92.7 | |
| SoftLexiconArchitecture=LSTM2021.07 | 93.66 | 94.63 | 92.7 | |
| LGN2021.07 | 93.46 | 94.19 | 92.73 | |
| PLT2021.07 | 93.26 | 94.25 | 92.3 | |
| Zhang and Yang [10]Data=External Lexicon2019.04 | 93.18 | 93.57 | 92.79 | |
| Lattice-LSTMArchitecture=Lattice-based, Sequence modeling=LSTM2019.08 | 93.18 | 93.57 | 92.79 | |
| Lattice-LSTM2021.07 | 93.18 | 93.57 | 92.79 | |
| Zhang and Yang (2018)2021.12 | 93.18 | 93.57 | 92.79 | |
| CAN Model2019.04 | 92.97 | 93.53 | 92.42 | |
| CAN-NER2021.07 | 92.97 | 93.53 | 92.42 | |
| Yan et al. (2019)2021.12 | 92.74 | — | — | |
| PRODEN-mlpMethod Category=PLL2023.05 | 92.46 | — | — | |
| Char-based (LSTM) + bichar + ExSoftwordArchitecture=Char-based, Sequence modeling=LSTM, Features=bichar+ExSoftword2019.08 | 92.38 | 93.21 | 91.57 | |
| Baseline + CNN2019.04 | 92.34 | 92.57 | 92.11 | |
| SeqcrowdMethod Category=Annotator2023.05 | 91.9 | — | — | |
| Char-based (LSTM) + bichar + softwordArchitecture=Char-based, Sequence modeling=LSTM, Features=bichar+softword2019.08 | 91.87 | 92.97 | 90.8 | |
| Zhou et al. [32]Architecture=CNN-BiLSTM-CRF, Features=Stroke embeddings and n-gram features2019.04 | 91.67 | 92.04 | 91.31 | |
| Zhang et al. [28]Features=Hand-crafted2019.04 | 91.18 | 92.2 | 90.18 | |
| Zhang et al., 2006*2019.08 | 91.18 | 92.2 | 90.18 | |
| Zhang et al. 2006*2019.08 | 91.18 | 92.2 | 90.18 | |
| Zhang et al. (2006)*2021.07 | 91.18 | 92.2 | 90.18 | |
| Dong et al. [31]Features=Radical features2019.04 | 90.95 | 91.28 | 90.62 | |
| Dong et al. 20162019.08 | 90.95 | 91.28 | 90.62 | |
| Dong et al. 20162019.08 | 90.95 | 91.28 | 90.62 | |
| Dong et al. (2016)2021.07 | 90.95 | 91.28 | 90.62 | |
| Cao et al. [11]Architecture=Adversarial Transfer Learning and Global self-attention2019.04 | 90.64 | 91.3 | 89.58 | |
| Baseline2019.04 | 90.32 | 92.54 | 88.2 | |
| Zhou et al. [29]Features=Hand-crafted2019.04 | 90.28 | 91.86 | 88.75 | |
| Zhou et al., 20132019.08 | 90.28 | 91.86 | 88.75 | |
| Zhou et al. 20132019.08 | 90.28 | 91.86 | 88.75 | |
| Zhou et al. (2013)2021.07 | 90.28 | 91.86 | 88.75 | |
| Char-based (LSTM) + ExSoftwordArchitecture=Char-based, Sequence modeling=LSTM, Features=ExSoftword2019.08 | 88.97 | 90.77 | 87.23 | |
| Char-based (LSTM)Architecture=Char-based, Sequence modeling=LSTM2019.08 | 88.81 | 90.74 | 86.96 | |
| Char-based (LSTM)2019.08 | 88.81 | 90.74 | 86.96 | |
| Lu et al. [30]Embeddings=Multi-prototype2019.04 | 87.94 | — | — | |
| Lu et al. 20162019.08 | 87.94 | — | — | |
| Lu et al. 20162019.08 | 87.94 | — | — | |
| Lu et al. (2016)2021.07 | 87.94 | — | — | |
| Chen et al. [27]Features=Hand-crafted2019.04 | 86.2 | 91.22 | 81.71 | |
| Chen et al., 20062019.08 | 86.2 | 91.22 | 81.71 | |
| Chen et al., 20062019.08 | 86.2 | 91.22 | 81.71 | |
| Chen et al. (2006)2021.07 | 86.2 | 91.22 | 81.71 |