Research on Natural Language Interface Method Based on Knowledge Graph Enhancement
Natural language inference(NLI)is an important task in natural language processing.It aims to identify the logical relationship that exists between two sentences.Most existing methods use the semantic knowledge obtained from the training corpus for reasoning.But it ignores the use and introduction of background knowledge.In this work,to solve this problem,a new NLI(KG-NET)model is proposed based on knowledge graph enhancement,so as to introduce the enhancement of related domain knowledge in the NLI task.The KGNET model is composed of three components,which are a semantic relationship representation module,a knowledge relationship representation module and a label prediction module.The experiments of the model on two benchmark datas-ets(SNLI and MultiNLI)verify the effectiveness of the model.
natural language interfaceknowledge graphgraph neural networkexternal knowledge
国家科技期刊平台
NSTL
万方数据
