bert-kbqa-NLPCC2017

March 10, 2019 ยท View on GitHub

A trial of kbqa based on bert for NLPCC2016/2017 Task 5

Step1: Get the entity from question with BERT-BiLSTM-CRF

Step2: Map the question to the sematically similartiy candidate attribute

Contents

  1. 1Step1: Get the entity from question with BERT-BiLSTM-CRF
  2. 2Step2: Map the question to the sematically similartiy candidate attribute