Automatic Question Answering Method Based on IMGRU-Seq2seq
Yujiao Jiang, Lvwen Huang, Zimeng Jia, Bin Yang · 2021 IEEE 6th International Conference on Computer and Communication Systems (ICCCS) · 2021
Aiming at the problem of gradient disappearance and network degradation caused by using of recurrent neural networks in traditional question answering models, an automatic question answering model based on IMGRU-Seq2seq is proposed. First, based on the gated recurrent unit, the batch normalization technology and the rectified linear unit activation function are combined and the identity mapping is added to construct the IMGRU model; then, the text is represented by weighted word vectors through the TF-IDF method; finally, as the semantic extraction unit of the question answering model, the bidirectional IMGRU introduced attention mechanism and the beam search algorithm to realize automatic question and answer. The experimental results showed that the proposed method is 16.55% and 8.36% higher than the existing methods BLEU and ROUGE-L respectively.