Knowledge graph embedding with self adaptive double-limited loss

Many well-performing embedding models for knowledge graphs employ a negative sampling framework to complete the