FINE-TUNING METHOD AND APPARATUS FOR EXTERNAL KNOWLEDGE-FUSING BERT MODEL, AND COMPUTER DEVICE

摘要

A fine-tuning method and apparatus for an external knowledge-fusing BERT model, and a computer device. The method comprises: upon receiving an inputted Chinese sentence, obtaining a sentence vector and part-of-speech vectors of the Chinese sentence according to a BERT model (S110); extracting from a preset external knowledge base a sememe set of the Chinese sentence (S120); inputting sememes of the sememe set into the BERT model to obtain a sememe vector set of the sememe set (S130); filtering out of the sememe vector set the sememe vectors of the Chinese sentence (S140); and, on the basis of a preset fusion rule, fusing the sentence vector, the part-of-speech vectors, and the sememe vectors of the Chinese sentence so as to fine-tune the BERT model (S150). The method is based on natural language processing technology in artificial intelligence. In the invention, external knowledge is fused into a BERT model so as to fine-tune said model, thereby increasing the text analysis accuracy of the BERT model, with said external knowledge further allowing for the expanded text analysis capability of the BERT model.