A hierarchical semantic extraction model for Chinese counseling question based on neural networks

Abstract

In the Chinese natural language processing, this paper proposes a hierarchical semantic learning model based on the neural network model for the semantic understanding of the counseling question. The attention mechanism is used to integrate the factual part of the original problem into the core question part and it can enrich the final representation of the semantic information and highlight the main features. Experiments show that the hierarchical learning structure can extract the structural features of the counseling question well, and contains more semantic information than the traditional structure which directly learns the problem, so that the final vector has higher similarity in the space.