WebNLG
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Joint Entity and Relation Extraction | WebNLG (test) | Precision94.6 | 52 | |
| Metrics correlation with human judgment | WebNLG challenge 2017 | Spearman Correlation (rho)0.9 | 45 | |
| Data-to-text generation | WebNLG (test) | BLEU64.11 | 39 | |
| Relation Triple Extraction | WebNLG original (test) | F1 Score (%)94.7 | 33 | |
| Text Generation | WebNLG seen categories (test) | BLEU63.69 | 18 | |
| Graph-to-text generation | WebNLG (test) | Fluency4.3 | 18 | |
| Natural Language Generation | WebNLG unseen categories | BLEU49.8 | 17 | |
| Table-to-text generation | WebNLG | BLEU (Seen)65.4 | 17 | |
| Natural Language Generation | WebNLG all categories (test) | BLEU55.27 | 13 | |
| Data-to-text generation | WebNLG en | ROUGE-255.52 | 12 | |
| Graph-to-text generation | WebNLG seen v1.0 (test) | BLEU65.05 | 12 | |
| Relational Triplet Extraction | WebNLG | Partial F177.7 | 11 | |
| Natural Language Generation | WebNLG | BLEU55.3 | 11 | |
| Natural Language Generation | WebNLG all categories | BLEU57.7 | 11 | |
| Natural Language Generation | WebNLG seen categories | BLEU65.3 | 11 | |
| Graph-to-text generation | WebNLG all v1.0 (test) | BLEU59.7 | 11 | |
| Relational Fact Extraction | WebNLG (test) | Partial Precision77 | 11 | |
| Graph-to-text generation | WebNLG unseen v1.0 (test) | BLEU53.67 | 10 | |
| Data-To-Text | GEM WebNLG ru | ROUGE-225.5 | 9 | |
| Graph-to-Text | WebNLG v2.0 (test) | BLEU66.2 | 9 | |
| Natural Language Generation | WebNLG en | ROUGE-253.5 | 9 | |
| Table-to-text generation | WebNLG Unseen (test) | BLEU52.8 | 9 | |
| Table-to-text generation | WebNLG Seen (test) | BLEU Score64.7 | 9 | |
| Data-to-text generation | WebNLG Unseen v1 | Fluency Score5.63 | 9 | |
| Data-to-text generation | WebNLG Seen v1 | BLEU57.2 | 9 |
Showing 25 of 62 rows