HyperAI
الرئيسيةالأخبارأحدث الأوراق البحثيةالدروسمجموعات البياناتالموسوعةSOTAنماذج LLMلوحة الأداء GPUالفعاليات
البحث
حول
HyperAI

Relation Extraction On Ade Corpus

المقاييس

NER Macro F1
RE+ Macro F1

النتائج

نتائج أداء النماذج المختلفة على هذا المعيار القياسي

اسم النموذج
NER Macro F1
RE+ Macro F1
Paper TitleRepository
multi-head86.4074.58Joint entity recognition and relation extraction as a multi-head selection problem
PFN (ALBERT XXL, average aggregation)91.583.9An Information Extraction Study: Take In Mind the Tokenization!
REBEL (including overlapping entities)-82.2REBEL: Relation Extraction By End-to-end Language generation
Deeper89.4883.74Deeper Task-Specificity Improves Joint Entity and Relation Extraction
SpERT.PL (with overlap and BioBERT)91.1782.03Joint Entity and Relation Extraction from Scientific Documents: Role of Linguistic Information and Entity Types
Relation-Metric87.0277.19Neural Metric Learning for Fast End-to-End Relation Extraction-
CMAN89.4081.14Modeling Dense Cross-Modal Interactions for Joint Entity-Relation Extraction-
SpERT (without overlap)89.2579.24Span-based Joint Entity and Relation Extraction with Transformer Pre-training
PFN (ALBERT XXL, no aggregation)91.383.2A Partition Filter Network for Joint Entity and Relation Extraction
SpERT (with overlap)89.2878.84Span-based Joint Entity and Relation Extraction with Transformer Pre-training
CLDR + CLNER88.379.97Imposing Relation Structure in Language-Model Embeddings Using Contrastive Learning
SpERT.PL (without overlap and BioBERT)91.1482.39Joint Entity and Relation Extraction from Scientific Documents: Role of Linguistic Information and Entity Types
Table-Sequence89.780.1Two are Better than One: Joint Entity and Relation Extraction with Table-Sequence Encoders
ITER92.63 ± 0.8985.6 ± 1.42ITER: Iterative Transformer-based Entity Recognition and Relation Extraction
multi-head + AT86.7375.52Adversarial training for multi-context joint entity and relation extraction
0 of 15 row(s) selected.