Pre-training language model construction method based on knowledge graph

The invention provides a pre-training language model construction method based on a knowledge graph, and the method comprises the steps: carrying out the data processing of an input natural language problem based on a semantic analysis model of a pre-training language, obtaining a candidate logic form list, and enabling the semantic analysis model to be used for extracting the features of the natural language problem, entity recognition and intention recognition are carried out on the question, and a candidate logic form list is generated through logic conversion filling in combination with entities and intentions; and performing knowledge retrieval on the generated candidate logic form list in the knowledge graph based on an unsupervised multi-stage search algorithm, updating the logic form content in combination with a retrieval result, converting the logic form content into a Cypher statement for graph query, and returning an answer list to form an answer set. According to the method, the intelligent quest