Joint Entity And Relation Extraction On

Metrics

Cross Sentence

Entity F1

Results

Performance results of various models on this benchmark

			Paper Title
SpERT.PL (SciBERT)	No	70.53	Joint Entity and Relation Extraction from Scientific Documents: Role of Linguistic Information and Entity Types
SpERT	No	70.33	Span-based Joint Entity and Relation Extraction with Transformer Pre-training
SpERT (with overlap)	No	70.3	Span-based Joint Entity and Relation Extraction with Transformer Pre-training
TriMF	-	70.17	A Trigger-Sense Memory Flow Framework for Joint Entity and Relation Extraction
PL-Marker	Yes	69.9	Packed Levitated Marker for Entity and Relation Extraction
Ours: cross-sentence	Yes	68.9	A Frustratingly Easy Approach for Entity and Relation Extraction
MGE	No	68.4	A Multi-Gate Encoder for Joint Entity and Relation Extraction
DyGIE++	Yes	67.50	Entity, Relation, and Event Extraction with Contextualized Span Representations
PFN	No	66.8	A Partition Filter Network for Joint Entity and Relation Extraction
DyGIE	Yes	65.2	A General Framework for Information Extraction using Dynamic Span Graphs
SciIE	No	64.20	Multi-Task Identification of Entities, Relations, and Coreference for Scientific Knowledge Graph Construction

0 of 11 row(s) selected.

Joint Entity And Relation Extraction On | SOTA | HyperAI