Natural Language Understanding on STREUSLE
Metric: Full F1 (Preps) (higher is better)
Metric: Full F1 (Preps) (higher is better)
| # | Model↕ | Full F1 (Preps)▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | RoBERTa + Linear | 78.2 | No | - | - | - |
| 2 | BERT (pred POS/lemmas) | 71.6 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 3 | BERT (gold POS/lemmas) | 71.4 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 4 | BERT (none) | 70.9 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 5 | GloVe (gold POS/lemmas) | 61 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 6 | SVM (feature-rich, gold syntax) | 59.5 | No | Comprehensive Supersense Disambiguation of Engli... | 2018-05-13 | Code |
| 7 | BiLSTM + MLP (gold syntax) | 58.9 | No | Comprehensive Supersense Disambiguation of Engli... | 2018-05-13 | Code |
| 8 | GloVe (none) | 58.1 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 9 | GloVe (pred POS/lemmas) | 58 | No | Lexical Semantic Recognition | 2020-04-30 | Code |
| 10 | SVM (feature-rich, auto syntax) | 55.7 | No | Comprehensive Supersense Disambiguation of Engli... | 2018-05-13 | Code |
| 11 | BiLSTM + MLP (auto syntax) | 53.6 | No | Comprehensive Supersense Disambiguation of Engli... | 2018-05-13 | Code |