Бенчмарки извлечения отношений

AbstractBi-encoder architectures for distantly-supervised relation extraction are designed to make use of the complementary information found in text and knowledge graphs (KG).However, current architectures suffer from two drawbacks. They either …...