Named Entity Recognition on Twitter NER
78.52F1 ScoreRoBERTa-BiLSTM-CRF + WOM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RoBERTa-BiLSTM-CRF + WOMEnhancement=Window-oriented Optimization Module2026.04 | 78.52 | — | |
| BERTBackbone=BERT-base2021.05 | 77.97 | 1 | |
| RoBERTa-BiLSTM-CRFArchitecture=BiLSTM-CRF, Backbone=RoBERTa2026.04 | 77.85 | — | |
| TOKEECompression=~2X2021.05 | 77.77 | 2 | |
| Context + WOMEnhancement=Window-oriented Optimization Module2026.04 | 77.47 | — | |
| SENTEECompression=~2X2021.05 | 77.15 | 2.03 | |
| Context(RoBERTa)Backbone=RoBERTa2026.04 | 76.7 | — | |
| MINER + WOMEnhancement=Window-oriented Optimization Module2026.04 | 76.24 | — | |
| SpanNER + WOMEnhancement=Window-oriented Optimization Module2026.04 | 75.94 | — | |
| E-SpanNERBackbone=BERT large2023.05 | 75.64 | — | |
| MINER(BERT)Backbone=BERT2026.04 | 75.45 | — | |
| E-SpanNERBackbone=BERT base2023.05 | 74.94 | — | |
| SpanNER(BERT)Backbone=BERT2026.04 | 74.93 | — | |
| EDL-SpanNERBackbone=BERT large2023.05 | 74.14 | — | |
| TOKEECompression=~4X2021.05 | 74.01 | 3.86 | |
| DataAug2023.05 | 73.69 | — | |
| BERT2026.04 | 72.61 | — | |
| SpanNERBackbone=ROBERTa large2023.05 | 71.7 | — | |
| SpanNERBackbone=BERT large2023.05 | 71.57 | — | |
| VaniIB2023.05 | 71.19 | — | |
| SpanNERBackbone=ALBERT large2023.05 | 70.33 | — | |
| SENTEECompression=~4X2021.05 | 66.47 | 4.14 | |
| LSTM-CRF2021.05 | 65.32 | — |