multi-head + AT

Reported on 6 benchmarks across 1 task · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing6 results

Relation ExtractiononAdverse Drug Events (ADE) Corpus
NER Macro F1· 2018-08-21
86.73
best: 91.5 (PFN (ALBERT XXL, average aggregation))
SOTA
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876
Relation ExtractiononAdverse Drug Events (ADE) Corpus
RE+ Macro F1· 2018-08-21
75.52
best: 83.9 (PFN (ALBERT XXL, average aggregation))
SOTA
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876
Relation ExtractiononACE 2004
NER Micro F1· 2018-08-21
81.64
best: 90.4 (PL-Marker)
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876
Relation ExtractiononACE 2004
RE+ Micro F1· 2018-08-21
47.45
best: 66.5 (PL-Marker)
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876
Relation ExtractiononCoNLL04
NER Macro F1· 2018-08-21
83.6
best: 87 (Deeper)
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876
Relation ExtractiononCoNLL04
RE+ Macro F1 · 2018-08-21
61.95
best: 76.65 (REBEL)
Adversarial training for multi-context joint entity and relation extraction arXiv:1808.06876