SEMANTIC CLASSIFICATION METHOD AND DEVICE, NEURAL NETWORK TRAINING METHOD AND DEVICE, AND MEDIUM

摘要

A semantic classification method and device, a neural network training method and device, and a storage medium. The semantic classification method comprises: inputting a first comment related to a first object; processing the first comment by using a common expression extractor, so as to extract a first common expression vector representing a common expression in the first comment; processing the first comment by using a first expression extractor, so as to extract a first individual expression vector used to represent an individual expression in the first comment; joining the first common expression vector and the first individual expression vector to obtain a first expression vector; and processing the first expression vector by using a first semantic classifier, so as to obtain a semantic classification result of the first comment, wherein the common expression comprises a meaning expression for performing commenting for both the first object and a second object, the second object is an associated comment object different from the first object, and the individual expression of the first comment comprises a meaning expression solely for performing commenting for the first object.