Event TEMPREL extraction on MATRES
85.7F1 ScoreMuLCo
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MuLCoPLM=BERT-base2026.03 | 85.7 | — | — | — | |
| WISTERIAPLM=BERT-base2026.03 | 84.3 | — | — | — | |
| CPTREPLM=BERT-base2026.03 | 84.2 | — | — | — | |
| Unified-FrameworkPretrained Model=RoBERTa-Large2023.05 | 82.6 | — | — | — | |
| Unified-FrameworkPLM=RoBERTa-Large2026.03 | 82.6 | — | — | — | |
| TIMERSPretrained Model=BERT-Large2023.05 | 82.3 | — | — | — | |
| TIMERSPLM=BERT-Large2026.03 | 82.3 | — | — | — | |
| Enhanced-BaselinePretrained Model=RoBERTa-Large2023.05 | 81.9 | — | — | — | |
| Time-EnhancedPretrained Model=RoBERTa-Large2023.05 | 81.7 | — | — | — | |
| SCS-EEREPretrained Model=RoBERTa-Large2023.05 | 81.6 | — | — | — | |
| HGRUPretrained Model=RoBERTa-Large2023.05 | 80.5 | — | — | — | |
| SyntacticPretrained Model=BERT-Large, additional training resources=true2023.05 | 80.3 | — | — | — | |
| Zhang et al. (2021)Protocol=Best single model2020.12 | 80.3 | — | — | — | |
| SyntacticPretrained Model=BERT-Base, additional training resources=true2023.05 | 79.3 | — | — | — | |
| Unified-FrameworkPretrained Model=BERT-Base2023.05 | 79.3 | — | — | — | |
| DEERPretrained Model=RoBERTa-Large2023.05 | 79.3 | — | — | — | |
| ECONETProtocol=Best single model, Backbone=RoBERTa-Large2020.12 | 79.3 | — | — | — | |
| HMHDPretrained Model=RoBERTa-Large2023.05 | 78.8 | — | — | — | |
| RoBERTa-Large + ECONETProtocol=Average of 3 seeds2020.12 | 78.8 | — | — | — | |
| Enhanced-BaselinePretrained Model=BERT-Base2023.05 | 78.5 | — | — | — | |
| RoBERTa-LargeProtocol=Average of 3 seeds2020.12 | 78.3 | — | — | — | |
| BoxPretrained Model=RoBERTa-Base2023.05 | 77.3 | — | — | — | |
| BERT-basedPLM=BERT-base2026.03 | 77.2 | — | — | — | |
| RoBERTa-Large + GeneratorProtocol=Average of 3 seeds2020.12 | 77 | — | — | — | |
| HNPPretrained Model=BERT-Base2023.05 | 75.5 | — | — | — | |
| LSTMPretrained Model=BERT-Base2023.05 | 73.4 | — | — | — | |
| CogCompTime2.0Pretrained Model=BERT-Base2023.05 | 71.4 | — | — | — | |
| TacoLMProtocol=Average of 3 seeds2020.12 | 70.9 | — | — | — | |
| BERT-LargeProtocol=Average of 3 seeds2020.12 | 70.5 | — | — | — | |
| BERT-Large + ECONETProtocol=Average of 3 seeds2020.12 | 70.4 | — | — | — | |
| BiLSTMPLM=word2vec2026.03 | 59.5 | — | — | — | |
| Faithful TEMPREL extractor (Ours)Full Model=true2022.10 | — | 82.7 | 56.3 | 3.4 | |
| Faithful TEMPREL extractor (Ours) w/o CACounterfactual Analysis=false2022.10 | — | 80.3 | 54.7 | 5 | |
| Faithful TEMPREL extractor (Ours) w/o DPDirichlet Prior=false2022.10 | — | 81.3 | 55.2 | 11.8 | |
| Faithful TEMPREL extractor (Ours) w/o TITense Information=false2022.10 | — | 81.8 | 55.2 | 2 | |
| Faithful TEMPREL extractor (Ours) w/o TSTemperature Scaling=false2022.10 | — | 82.6 | 56.1 | 49.6 | |
| Mathur et al. (2021)2022.10 | — | 82.3 | 55.7 | 12.8 | |
| Trong et al. (2022)2022.10 | — | 83.4 | 56.4 | 13 |