Research on Question Answering System Based on Bi-LSTM and Self-attention Mechanism

Abstract

With the development of artificial intelligence technology, intelligent question an-swering has become a hot research direction in the field of natural language pro-cessing. This paper proposes a question answering method based on Bi-LSTM and self-attention mechanism model. This method uses Bi-LSTM to encode and align the question and answer respectively, then uses self-attention to obtain the relationship between keywords, and finally performs softmax through the fully connected layer to obtain the similarity between the question and answer. Finally, in the experiment, compared with the traditional attention model, the accuracy rate of this model was increased by 1.6%, and the recall rate was increased by 1.5%.