A BERT-powered multi-head selection model with soft label embedding and weak-supervision NER pretraining reaches F1 0.876 single and 0.892 ensembled on the SKE Chinese information extraction benchmark.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
BERT-Based Multi-Head Selection for Joint Entity-Relation Extraction
A BERT-powered multi-head selection model with soft label embedding and weak-supervision NER pretraining reaches F1 0.876 single and 0.892 ensembled on the SKE Chinese information extraction benchmark.