PROBABILISTIC GRAPHICAL MODEL-BASED TEXT ATTRIBUTE EXTRACTION METHOD AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM

摘要

A probabilistic graphical model-based text attribute extraction method and apparatus, a computer device, and a storage medium, relating to artificial intelligence neural network technology. The method comprises: inputting into a BERT neural network model a received text to be processed, and obtaining corresponding text representation output; inputting the text representation output into a multi-task learning classification model so as to obtain a corresponding entity type; sequentially performing recursion, vector concatenation, feature fusion and essential-attribute extraction on the entity type so as to obtain the essential attributes in the entity and start and end positions of the essential attributes; and sequentially performing entity representation vector extraction, vector concatenation and feature fusion, and non-essential-attribute extraction on the essential attributes and the start and end positions of the essential attributes, so as to obtain non-essential attributes in the entity and start and end positions of the non-essential attributes. The invention improves the accuracy of attribute extraction from data. Furthermore, there are no data format restrictions on text to be processed; thus, any structured data or unstructured data may be inputted.